JPH09218938A - Image processing method and apparatus - Google Patents

Abstract

(57) Abstract: A block selection process and a character recognition process included in an image processing device cannot be instructed from another device. A local area network (LA)
N) and the interface unit 7 for performing data communication, and the external device 3 receives information for instructing execution of block selection processing and character recognition processing from another device connected to the LAN via the interface unit 7. When it is received, the image information is read from the reader unit 1, the core unit 10 performs image processing, and the image memory unit 8 or the printer unit 2
The processing result is output to.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing method and apparatus.

[0002]

2. Description of the Related Art Conventionally, block selection processing and character recognition processing of an image processing apparatus can be used and executed only in that image processing apparatus.

[0003]

Therefore, in order to execute the block selection process or the character recognition process by using the block selection process or the character recognition process included in the image processing apparatus, the execution is performed by directly operating the device. I had to give instructions.

Further, the block selection process and the character recognition process cannot be executed in other devices which are connected by the same communication line and have neither the block selection process nor the character recognition process.

The present invention has been made to solve the above-mentioned problems, and provides an image processing method and apparatus capable of instructing execution of predetermined image processing from another apparatus connected to a network through an interface. The purpose is to do.

[0006]

In order to achieve the above object, the image processing method of the present invention has the following steps.

That is, a receiving step of receiving instruction information for instructing execution of predetermined image processing from another device connected to the network via the interface, and based on the instruction information received by the receiving step, An input step of inputting image information, an image processing step of processing the image information, and a control step of controlling the input image information to perform predetermined image processing, and a predetermined step by the control step. And a transmitting step of transmitting the image on which the image processing is performed.

In order to achieve the above object, the image processing apparatus according to the present invention has the following configuration.

That is, interface means for connecting to a network, receiving means for receiving instruction information for instructing execution of predetermined image processing from another device connected to the network by the interface means,
Based on the instruction information received by the receiving means, the input means for inputting the image information and the image processing means for processing the image information are controlled so as to execute predetermined image processing on the input image information. And a transmission unit for transmitting the image on which the predetermined image processing has been performed by the control unit.

In such a configuration, the instruction information for instructing the execution of the image processing is received from the other device connected to the network via the interface, and based on the instruction information, the input of the image information and the image information thereof are performed. Control is performed so that predetermined image processing is executed, and the image on which the predetermined image processing is executed is transmitted.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT A preferred embodiment of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a block diagram showing the arrangement of an image processing apparatus according to this embodiment. In FIG. 1, reference numeral 1 is an image input device (hereinafter referred to as "reader unit"), which optically reads a document and converts it into image data. An image output device (hereinafter referred to as a "printer unit") 2 has a plurality of types of recording paper cassettes and outputs image data as a visible image on the recording paper according to a print command. An external device 3 is electrically connected to the reader unit 1 and has various functions. The external device 3 communicates with the fax unit 4, the file unit 5, the external storage device 6 connected to the file unit 5, and the local area network (LAN) as a circuit that executes various functions. In order to temporarily store the information from the LAN interface unit 7 and the reader unit 1 or the information sent from the computer, and to classify the attribute of the partial area in the image with respect to the temporarily stored image information. Processing (hereinafter referred to as block selection processing) and an image memory section 8 for performing various image processings, and a core section 1 for controlling the above-mentioned functions.
0 and a hard disk 11. 12 is a telephone line.

The above-mentioned parts will be described below in order.

<Description of Reader Unit 1> FIG. 2 shows the reader unit 1.
3 is a cross-sectional view showing the configuration of the printer unit 2 and the configuration and operation thereof will be described below.

The originals stacked on the original feeding device 101 are sequentially conveyed one by one onto the original table glass surface 102.
When the document is conveyed, the lamp 103 of the scanner unit is turned on, and the scanner unit 104 moves to irradiate the document. The reflected light from the document is reflected by the mirrors 105, 106,
After passing through the converging lens 108 via 107 in sequence, the CCD
Image sensor unit 109 (hereinafter referred to as "CCD")
Is input to

FIG. 3 is a circuit block diagram showing the signal processing configuration of the reader unit 1 described above, and the configuration and operation will be described below.

First, the image information input to the CCD 109 is photoelectrically converted here and converted into an electric signal. CCD
The color information (RGB) from 109 is output to the next amplifier 11
It is amplified by 0R, 110G, and 110B according to the input signal level of the A / D converter 111. A / D converter 111
The output signal from is input to the shading circuit 112, where light distribution unevenness of the lamp 103, sensitivity unevenness of the CCD, and the like are corrected. The signal from the shading circuit 112 is output to the Y signal generation / color detection circuit 113 and the external I / O circuit.
It is input to the F switching circuit 119, respectively.

The Y signal generation / color detection circuit 113 calculates the signal from the shading circuit 112 by the following equation to obtain the Y signal.

Y = 0.3R + 0.6G + 0.1B Further, it has a color detection circuit which separates each signal of R, G and B into seven colors and outputs a signal for each color. An output signal from the Y signal generation / color detection circuit 113 is input to a scaling / repeat circuit 114. Scaling in the sub-scanning direction is performed according to the scanning speed of the scanner unit 104, and scaling in the main scanning direction is performed by the scaling / repeat circuit 114. Further, the scaling / repeat circuit 114 can output a plurality of identical images. The contour / edge enhancement circuit 115
Edge emphasis and contour information are obtained by emphasizing high frequency components of the signal from the scaling / repeat circuit 114. The signal from the contour / edge emphasis circuit 115 is input to the marker area determination / contour generation circuit 116 and the patterning / thickening / masking / trimming circuit 117.

Marker area determination / contour generation circuit 116
Reads the portion of the manuscript written with a marker pen of a specified color and generates the contour information of the marker, and the next patterning / thickening / masking / trimming circuit 117 thickens, masks or trims the contour information. I do. Further, the output signal from the patterning / thickening / masking / trimming circuit 117 is output by the laser driver circuit 118 according to the color detection signal from the Y signal generation / color detection circuit 113.
The various processed signals input to are converted into signals for driving the laser. The output signal of the laser driver 118 is input to the printer unit 2 and image formation is performed as a visible image.

Next, the external I / F switching circuit 119 for interfacing with the external device 3 will be described.

First, when the external I / F switching circuit 119 outputs image information from the reader unit 1 to the external device 3, a patterning / thickening / masking / trimming circuit 117 is provided.
Is output to the connector 120. When the image information from the external device 3 is input to the reader unit 1, the external I / F switching circuit 119 inputs the image information from the connector 120 to the Y signal generation / color detection circuit 113.

The above-mentioned image processing is performed in accordance with an instruction from the CPU 122, and the area signal generation circuit 121 generates various timing signals necessary for the above-mentioned image processing according to the value set by the CPU 122. Further CPU12
The communication with the external device 3 is performed using the communication function built in the communication device 2. The SUB / CPU 123 communicates between the operation unit 124 and the external device 3 by using the built-in communication function.

<Description of Printer Unit 2> Referring back to FIG. 2, the configuration and operation of the printer unit 2 will be described.

The signal input to the printer unit 2 is converted into an optical signal by the exposure control unit 201, and the photoconductor 202 is illuminated according to the image signal. The latent image formed on the photoconductor 202 by the irradiation light is developed by the developing device 203. The transfer paper is conveyed from the transfer paper stacking section 204 or 205 at the same timing as the development, and the developed image is transferred to the transfer section 206. The transferred image is fixed on the transfer target paper by the fixing unit 207, and then the paper output unit 2
08 to the outside of the apparatus. The transfer paper output from the paper output unit 208 is discharged to each bin when the sort function is working in the sorter 220 or to the highest bin of the sorter when the sort function is not working. .

Next, a method for outputting sequentially read images on both sides of one output sheet will be described.

After the output sheet fixed by the fixing section 207 is once conveyed to the sheet discharge section 208, the conveying direction of the sheet is reversed, and the sheet is transferred to the re-transferred transfer sheet stacking section 210 via the conveying direction switching member 209. Transport. When the next manuscript is prepared,
The original image is read in the same manner as in the above process, but the transfer paper is fed from the re-feeding transfer target paper stacking section 210, so that two sheets are printed on the front and back sides of the same output paper.
It is possible to output one original image.

<Description of External Device> The external device 3 is connected to the reader unit 1 by a cable, and the core unit 10 in the external device 3 is connected.
Controls signals and controls each function. In the external device 3, a fax unit 4 for transmitting and receiving a fax, a file unit 5 for converting various document information into electric signals and storing the same, a LAN interface unit 7 for connecting to a LAN and performing data communication via the LAN, The information sent from the reader unit 1, the fax unit 4, the file unit 5, and the LAN interface unit 7 is temporarily stored, and the attribute of the partial area in the image is classified with respect to the temporarily stored image information. Image memory unit 8 for performing various image processing such as block selection processing, character recognition processing, and binarization processing, a core unit 10 for controlling the above-described functions, and a hard disk 11.
Contains.

Here, the function of each unit included in the external device 3 will be described in detail.

<Description of Core Unit> First, the detailed structure of the core unit 10 will be described below with reference to FIG.

The connector 1001 of the core unit 10 is connected to the connector 120 of the reader unit 1 by a cable. Four types of signals are input to and output from the connector 1001, and the signal 1057 is an 8-bit multi-value video signal. The signal 1055 is a control signal for controlling the video signal. The signal 1051 communicates with the CPU 122 in the reader unit 1.
The signal 1052 communicates with the SUB / CPU 123 in the reader unit 1. The signals 1051 and 1052 are the communication I
Communication protocol processing is performed in C1002, and CPU bus 10
Communication information is transmitted to the CPU 1003 via 53.

The signal 1057 is a bidirectional video signal, and is a signal exchanged between the reader unit 1 and the core unit 10. The signal 1057 is input to the buffer 1010 where it is converted from a bidirectional signal to a unidirectional signal 1058,10.
It is separated into 70. The signal 1058 is 8 from the reader unit 1.
Bit multi-valued video signal, LUT 1011 in the next stage
Is input to The LUT 1011 converts the image information from the reader unit 1 into a desired value using a look-up table. The output signal 1059 from the LUT 1011 is input to the binarization circuit 1012 or the selector 1013. 2
The binarization circuit 1012 has a simple binarization function for binarizing a multi-valued signal 1059 at a fixed slice level, and a binarization function for a variable slice level in which the slice level varies from the values of pixels around a target pixel. , And the error diffusion method 2
It has a valuation function. The binarized information is converted into a multilevel signal of 00H when it is “0” and FFH when it is “1”,
It is input to the selector 1013 in the next stage.

The selector 1013 selects either the signal from the LUT 1011 or the output signal of the binarization circuit 1012.
The output signal 1060 from the selector 1013 is input to the selector 1014. The selector 1014 includes a fax unit 4, a file unit 5, a LAN interface unit 7,
The output video signal from the image memory unit 8 is input to the core unit 10 via the connectors 1005, 1006, 1007 and 1008, respectively, or the signal 1064 or the selector 1 is input.
The output signal 1060 of 013 is selected by the instruction of the CPU 1003. The output signal 1061 of the selector 1014 is input to the rotation circuit 1015 or the selector 1016. The rotation circuit 1015 converts the input image signal into +90 degrees and -90 degrees.
It has the function of rotating by +180 degrees. Rotating circuit 10
Reference numeral 15 is a binarization circuit 10 for converting the information output from the reader unit 1.
After being converted into a binary signal at 12, it is stored in the rotating circuit 1015 as information from the reader unit 1. Next, according to an instruction from the CPU 1003, the rotation circuit 1015 rotates and reads the stored information. The selector 1016 outputs the output signal 1062 from the rotating circuit 1015 and the input signal 1 of the rotating circuit 1015.
061 is selected, and as the signal 1063, the connector 1005 to the fax unit 4 and the connector 1006 to the file unit 5 and the connector 1 to the LAN interface unit 7 are selected.
007, input to connector 1008 of image memory unit 8 and selector 1017.

A signal 1063 is a signal on a synchronous 8-bit unidirectional video bus for transferring image information from the core unit 10 to the fax unit 4, file unit 5, LAN interface unit 7, and image memory unit 8. A signal 1064 is a signal on a synchronous 8-bit unidirectional video bus for transferring image information from the fax unit 4, file unit 5, LAN interface unit 7, and image memory unit 8, respectively. The video control circuit 1004 controls the synchronous bus.
Control is performed by the output signal 1056 from 04. Other signals 1054 are provided to the connectors 1005 to 1008.
Are respectively connected. The signal 1054 is a bidirectional 16-bit CPU bus signal for exchanging data commands asynchronously. Information can be transferred between the fax unit 4, the file unit 5, the LAN interface unit 7, the image memory unit 8 and the core unit 10 by the above two video buses and the CPU bus.

A signal 10 from the fax unit 4, file unit 5, LAN interface unit 7, and image memory unit 8.
64 is input to the selector 1014 and the selector 1017. The selector 1016 outputs the signal 1064 to the rotation circuit 1015 at the next stage according to an instruction from the CPU 1003.

The selector 1017 outputs the signal 1063 or 10
64 is selected by the instruction of the CPU 1003. The output signal 1065 of the selector 1017 is the pattern matching 10
18 and the selector 1019. The pattern matching 1018 pattern-matches the input signal 1065 with a predetermined pattern, and when the patterns match, outputs a predetermined multi-valued signal to the signal line 1066. If they do not match in the pattern matching, the input signal 1065 is output as the signal 1066.

The selector 1019 outputs the signal 1064 or 10
66 is selected by the instruction of the CPU 1003. The output signal 1067 of the selector 1019 is input to the LUT 1020 of the next stage.

The LUT 1020 uses the input signal 1 in accordance with the characteristics of the printer when outputting the image information to the printer 2.
Convert 067. The selector 1021 is connected to the LUT 102
The output signal 1068 or 1065 of 0 is selected by the instruction of the CPU 1003. The output signal of the selector 1021 is input to the enlargement circuit 1022 of the next stage.

The enlarging circuit 1022 can set the enlarging magnification independently in the X and Y directions according to an instruction from the CPU 1003. The expansion method is a linear interpolation method of the first order. The output signal 1070 of the enlargement circuit 1022 is input to the buffer 1010.

The signal 107 input to the buffer 1010
0 is a bidirectional signal 1057 according to an instruction from the CPU 1003.
Then, it is sent to the printer 2 via the connector 1001 and printed out.

The flow of signals between the core section 10 and each section will be described below.

[Operation with Fax Unit 4] First, the core unit 1
A case where information is output from 0 to the fax unit 4 will be described. The CPU 1003 of the core unit 10 is the communication IC 100.
It communicates with the CPU 122 of the reader unit 1 via 2 and issues a document scan command. By this command, the scanner unit 104 of the reader unit 1 scans the document and outputs the image information to the connector 120. Reader unit 1 and external device 3
Are connected by a cable, and the information from the reader unit 1 is input to the connector 1001 of the core unit 10. Next, the information input to the connector 1001 is input to the buffer 1010 as a multi-valued 8-bit signal 1057. The buffer 1010 outputs the bidirectional signal 1057 as a unidirectional signal 1058 to the LUT 1011 according to an instruction from the CPU 1003. In this LUT 1011, the image information from the reader unit 1 is converted into a desired value using a look-up table. For example, it is possible to skip the background of a document.

Output signal 1059 from LUT 1011
Is input to the binarization circuit 1012 at the next stage. The binarization circuit 1012 converts the 8-bit multilevel signal 1059 into a binarized signal. The binarization circuit 1012 converts the binarized signal into two multi-level signals of 00H when it is “0” and FFH when it is “1”. The converted signal is sent to the selector 10
It is input to the rotation circuit 1015 or the selector 1016 via 13, 1014. Output signal 1 of rotating circuit 1015
062 is also input to the selector 1016, and the selector 101
6 selects either signal 1061 or 1062. The signal selection is determined by the CPU 1003 communicating with the fax unit 4 via the CPU bus. The output signal 1063 from the selector 1016 is the connector 100
It is sent to the fax unit 4 via 5.

Next, a case where information from the fax unit 4 is received and output to the printer unit 2 will be described.

First, the image information from the fax unit 4 is transmitted as a signal 1064 via the connector 1005.
The signal 1064 is input to the selectors 1014 and 1017. When the image at the time of fax reception is rotated and output to the printer unit 2 according to the instruction of the CPU 1003, the signal 1064 input to the selector 1014 is output to the rotation circuit 1015.
Rotate with. Output signal 10 from rotating circuit 1015
Reference numeral 62 designates selectors 1016 and 1017 and an expansion circuit 102.
5 is input to the pattern matching 1018.

The printer unit 2 is instructed by the CPU 1003.
If you want to rotate and output the image at the time of fax reception,
The signal 1064 input to the selector 1014 is used as the expansion circuit 1
At 015, a rotation process is performed. The output signal 1062 from the rotation circuit 1015 is input to the pattern matching 1018 via the selectors 1016 and 1017.

The pattern matching 1018 has a function of smoothing the rattling of the image when the fax is received. It is input to the LUT 1020 via the selector 1019. The LUT 1020 is used to output the image received by fax to the printer unit 2 at a desired density.
The table 020 is configured to be changeable by the CPU 1003. Output signal 1068 of LUT 1020
Is input to the enlargement circuit 1022 via the selector 1021. The expansion circuit 1022 has two values (00H, FF
The 8-bit multivalue having H) is enlarged by the linear interpolation method of the first order. The 8-bit multi-valued signal with many values from the expansion circuit 1022 is stored in the buffer 1010 and the connector 1.
It is sent to the reader unit 1 via 001. As a result, the reader unit 1 sends this signal to the external I via the connector 120.
It is input to the / F switching circuit 119. The external I / F switching circuit 119 generates a Y signal from the signal from the fax unit 4.
It is input to the color detection circuit 113. The output signal from the Y signal generation / color detection circuit 113 is subjected to the above-described processing and then output to the printer unit 2 to form an image on an output sheet.

[Operation with File Unit 5] First, the case where the information from the reader unit 1 is output to the file unit 5 will be described. The CPU 1003 communicates with the CPU 122 of the reader unit 1 via the communication IC 1002 and issues a document scan command. According to this command, the scanner unit 104 of the reader unit 2 scans the document, and outputs image information to the connector 120. Reader unit 1 and external device 3
Are connected by a cable, and the information from the reader unit 1 is input to the connector 1001 of the core unit 10. The image information input to the connector 1001 becomes a unidirectional signal 1058 by the buffer 1010. The signal 1058 which is a multi-valued 8-bit signal is converted into a desired signal by the LUT 1011. Output signal 1059 of LUT 1011
Is input to the connector 1006 via the selectors 1013, 1014 and 1016.

That is, the binarization circuit 1012 and the rotation circuit 1
The 8-bit multi-value data is transferred to the file unit 5 without using the function of 015. When filing a binarized signal by communicating with the file unit 5 via the CPU bus of the CPU 1003, the binarization circuit 1012, the rotation circuit 101
Use the function of 5. The binarization process and the rotation process are the same as those in the fax unit 4 described above, and therefore will be omitted.

Next, the information is received from the file section 5,
The case of outputting to the printer unit 2 will be described. The image information from the file unit 5 is input to the selector 1014 or 1017 as a signal 1064 via the connector 1006. Here, it is possible to input to the selector 1017 in the case of 8-bit multi-valued filing and to the selector 1014 or 1017 in the case of binary filing.
In the case of binary filing, the description is omitted because it is the same processing as the fax unit 4.

In the case of multi-value filing, the output signal 1065 from the selector 1017 is output to the LUT 1003 via the selector 1019. The LUT 1020 creates a look-up table according to an instruction from the CPU 1003 in accordance with a desired print density. LUT102
The output signal 1068 from 0 is input to the expansion circuit 1022 via the selector 1021. 8-bit multilevel signal 10 expanded to a desired expansion ratio by the expansion circuit 1022
70 is sent to the reader unit 1 via the buffer 1010 and the connector 1001. Then, the information of the file unit 5 sent to the reader unit 1 is output to the printer unit 2 and image formation is performed on the output paper, as in the above-described fax unit 4.

[Operation with LAN interface section 7]
First, a case where the image information read by the reader unit 1 is output to the LAN interface unit 7 will be described.
The CPU 1003 communicates with the CPU 122 of the reader unit 1 via the communication IC 1002 and issues a document scan command. The reader unit 1 outputs the image information to the connector 120 when the scanner unit 104 scans the document according to this command. The reader unit 1 and the external device 3 are connected by a cable, and information from the reader unit 1 is input to the connector 1001 of the core unit 10. Connector 10
The image information input to 01 becomes a unidirectional signal 1058 by the buffer 1010. The signal 1058, which is a multi-valued 8-bit signal, is sent to the LUN 1 via the selector 1023.
It is input to 011 and converted into a desired signal here.
The output signal 1059 of the LUT 1011 is the selector 101
3, input to the connector 1007 via the selector 1014 and the selector 1016, and multi-valued image information is transferred from there to the LAN interface unit 7.

Next, a case where information is received from the LAN interface unit 7 and output to the printer unit 2 will be described. Image information from the LAN interface unit 7 is input as a signal 1064 to the selector 1014 or the selector 1017 via the connector 1007. When the image information is 8-bit multivalued, it can be input to the selector 1017, and when it is binary, it can be input to the selector 1014 or 1017. If the image is a binary image, the description thereof is omitted because it is the same as that of the fax. Further, in the case of a multi-valued image, since it is the same as the case of multi-valued filing in the file part, description thereof will be omitted.

[Operation with Image Memory Unit 8] First,
A case where the image information read by the reader unit 1 is output to the image memory unit 8 will be described. CPU 100
3 is a CPU 12 of the reader unit 1 via the communication IC 1002
2 and communicates with 2 to issue a document scan command. In response to this command, the scanner unit 104 of the reader unit 1 scans the document to output the image information to the connector 120. The reader unit 1 and the external device 3 are connected by a cable, and the image information from the reader unit 1 is input to the connector 1001 of the core unit 10. The image information input to the connector 1001 is sent to the LUT 1011 via the buffer 1010 as a multilevel 8-bit signal 1057.
The output signal 1059 from the LUT 1011 is the selector 10
It is transferred to the image memory unit 8 as a multi-valued 8-bit image or a binarized image via 13, 10, 14 and 1016 and the connector 1008. Since the binarization process is the same as that of the fax unit 4, description thereof will be omitted.

Next, the case where the image information from the fax unit 4 is output to the image memory unit 8 will be described. The image information from the fax unit 4 is sent via the connector 1005.
It is transmitted as a multi-valued signal 1064 having one value (00H, FFH). The signal 1064 is input to the selector 1014. The signal transmitted to the selector 1014 is input to the rotation circuit 1015 or the selector 1016. The output signal 1062 of the rotation circuit 1015 is also input to the selector 1016, and the selector 1016 outputs the signal 1061 or the signal 106.
Any one of 62 is selected. The signal is selected by the CPU 100
3 communicates with the fax unit 4 via the CPU bus to determine it. The output signal 1063 from the selector 1016 is transferred to the image memory unit 8 via the connector 1008.

Next, the case of outputting the image information from the file section 5 to the image memory section 8 will be described. The image information from the file unit 5 is transmitted as a signal 1064 via the connector 1006. In the case of a multi-valued 8-bit image, the signal 1023 is sent to the LUT 1011 and in the case of a binary image, two values (00H, FFH). Is input to the selector 1014 as a multi-valued signal. LUT101
After the signal input to 1 is converted to the desired signal,
2 via the binarization circuit 1012 by the selector 1013
It is also possible to perform a digitization process. Since the binarization processing has been described in the fax unit 4, it will be omitted here. The signal transmitted to the selector 1014 is input to the rotation circuit 1015 or the selector 1016. The output signal 1062 of the rotation circuit 1015 is also input to the selector 1016, and the selector 1016 selects either the signal 1061 or the signal 1062. The signal selection is determined by the CPU 1003 communicating with the file unit 5 via the CPU bus. Then, the output signal 1063 from the selector 1016
Are transferred to the image memory unit 8 via the connector 1008.

Next, the case of outputting the data of the LAN interface section 7 to the image memory section 8 will be described. Data from the LAN interface unit 7 is transmitted as a signal 1064 via the connector 1007. If the data is a multi-valued 8-bit image, LUT10
11 and two values (0
0H, FFH) as a multilevel signal having a selector 101
4 is input. Further, even when the data is data other than image data, it is transmitted to the LUT 1011. LUT1
The multi-valued image signal transmitted to 011 is converted into a desired signal and then input to the selector 1013. Here, the selector 1013 outputs 2 through the binarization circuit 1012.
It is also possible to perform a digitization process. The details of the binarization have been described in the fax unit 4 and will be omitted here.

Further, in the case of data other than image data, the conversion in the LUT 1011 and the binarization in the binarization circuit 1012 are not performed, and the data is input to the selector 1013. The signal from the selector 1013 is further input to the rotation circuit 1015 or the selector 1016 via the selector 1014. The output signal 1062 of the rotation circuit 1015 is also the selector 1016.
Is input to However, the multi-valued 8-bit image and the data other than the image are output to the selector 1016 from the signal line 1061 without passing through the rotation circuit 1015. The selector 1016 selects either the signal 1061 or the signal 1062. When selecting the signal, the CPU 1003 executes CP
It is determined by communicating with the LAN interface unit 7 via the U bus 1054. The output signal 1063 from the selector 1016 is transferred to the image memory unit 8 via the connector 1008.

Next, the case where the image information from the image memory unit 8 is output from the printer unit 2 will be described. The image memory unit 8 transmits the 8-bit multilevel signal 1064 to the selectors 1014 and 1017 via the connector 1008. The output signal from the selector 1014 or 1017 is instructed by the CPU 1003, and the above-mentioned file unit 5
Similarly to the above, the image is output to the printer unit 2 and an image is formed on the output paper.

Next, the case where the image information from the image memory unit 8 is output from the fax unit 4 will be described. The image memory unit 8 outputs the image signal 1064 via the connector 1008 to the LUT when the image signal 1064 is 8-bit multivalued image information.
In the case of 1011, the selector 10
14 is transmitted. The multivalued image information transmitted to the LUT 1011 is binarized by the binarization circuit 1012 and then input to the selector 1014. Fax section 4 for details of binarization
Since it was mentioned above, it is omitted here.

The signal transmitted to the selector 1014 is input to the rotation circuit 1015 or the selector 1016. The output signal 1062 of the rotation circuit 1015 is also input to the selector 1016, and the selector 1016 outputs the signal 1061 or 106.
Select either of the two. The signal is selected by the CPU 1003.
Is determined by communicating with the fax unit 4 via the CPU bus. The output signal 1063 from the selector 1016 is sent to the fax unit 4 via the connector 1005.
Sent to

Next, the case where the image information from the image memory unit 8 is output from the file unit 5 will be described. The image memory unit 8 outputs the image signal 1064 via the connector 1008 to the LUT 1011 in the case of a multi-valued 8-bit image.
In the case of a binary image, two values (00
H, FFH) as a multilevel signal having a selector 1014
To be transmitted. The multi-valued image signal transmitted to the LUT 1011 is converted into a desired signal and then the selector 1013
Is input to Here, it is also possible to perform binarization processing by the selector 1013 via the binarization circuit 1012. The details of the binarization have been described in the fax unit 4 and will be omitted here.

The signal transmitted to the selector 1014 is input to the rotating circuit 1015 or the selector 1016. The output signal 1062 of the rotation circuit 1015 is also input to the selector 1016, and the selector 1016 outputs the signal 1061 or 106.
Select either of the two. The signal is selected by the CPU 1003.
Is determined by communicating with the file unit 5 via the CPU bus. The output signal 1063 from the selector 1016 is sent to the file unit 5 via the connector 1006.
Sent to

Next, the case where the data from the image memory unit 8 is output from the LAN interface unit 7 will be described. First, the image memory unit 8 includes the connector 100
If the data signal 1064 is a multi-valued 8-bit image or data other than image information, a LUT
In the case of a binary image, the selector 1 is transmitted as a multi-valued signal having two values (00H, FFH).
014. The multi-valued signal input to the LUT 1011 is converted into a desired signal by the LUT 1011 and then
It is input to the selector 1013. Here selector 101
It is also possible to perform binarization processing via the binarization circuit 1012 according to No. 3. The details of the binarization are described in the fax, and are omitted here.

The signal from the selector 1013 is further transmitted through the selector 1014 to the rotation circuit 1015 or the selector 10
16 is input. Output signal 106 of rotating circuit 1015
2 is also input to the selector 1016, and the selector 1016
Selects either the signal 1061 or the signal 1062. The signal is selected by the CPU 1003 on the CPU bus 105.
It is determined by communicating with the LAN interface unit 7 via the communication interface 4. Output signal 1 from the selector 1016
063 is sent to the LAN interface unit 7 via the connector 1006.

<Explanation of Fax Unit 4> Next, the detailed configuration of the fax unit 4 will be described below with reference to FIG.

FIG. 5 is a block diagram showing a detailed configuration of the fax unit 4. The fax unit 4 is connected to the core unit 10 by the connector 400 and exchanges various signals. The binary information from the core unit 10 is stored in the memory A405 to the memory D4.
08, the signal 453 from the connector 400 is input to the memory controller 404, and under the control thereof, the memory A 405, the memory B 406, and the memory C.
407, the memory D408, or two sets of memories connected in cascade. The memory controller 404 operates according to an instruction from the CPU 412 and stores the memory A40.
5, memory B406, memory C407, memory D408
, A mode for exchanging data with the CPU bus 462, a mode for exchanging data with the CODEC bus 463 of the CODEC 411 having an encoding and decoding function, a memory A405, a memory B406, and a memory C40.
7, the contents of the memory D408 the DMA controller 402
And a timing generation circuit 409 in which data is exchanged with the bus 454 from the scaling circuit 403 under the control of
The binary video input data 453 under the control of the memory A4
05-memory D408 has a function of storing in any one of them, and a function of reading the memory contents from any of the memories A405-D408 and outputting as a signal 452. In addition, the memory A405, the memory B406,
The memory C407 and the memory D408 each have a capacity of 2 Mbytes and store an image corresponding to A4 at a resolution of 400 dpi.

The timing generation circuit 409 is connected to the connector 4
00, and the control signal (H
SYNC, HEN, VSYNC, VEN) 459 and signal 46 to achieve the following two functions:
Generate 0. The first is a function of storing the image signal from the core unit 10 in any one of the memories A405 to D408 or two memories, and the second is the memory A.
A function of reading an image signal from any one of the memory 405 to the memory D408 and transmitting it as a signal 452. The dual port memory 410 exchanges signals 461 with the CPU 1003 of the core unit 10 via the connector 400, and is connected to the CPU 412 of the fax unit 4 via an internal bus. Each CPU exchanges commands via the dual port memory 410. SCSI
The controller 413 interfaces with the hard disk 11 connected to the fax unit 4 shown in FIG. Stores data such as fax transmission and reception data.

The CODEC 411 reads out the image information stored in any one of the memories A405 to D408, encodes the image information by a desired one of the MH, MR, and MMR systems, and then stores it in any one of the memories A405 to D408. Is stored as encoded information. Also, the memory A40
5 to read the coded information stored in the memory D408 and decode the coded information by a desired method of MH, MR, and MMR, and then decode the decoded information, that is, image information in any of the memories A405 to D408. Memorize as. M
The ODEM 414 has a function of modulating the coded information from the hard disk 11 connected to the CODEC 411 or the SCSI controller 413 for transmission on the telephone line, and demodulating the information sent from the NCU 415 to convert the coded information into coded information. Then, the encoded information is transferred to the hard disk 11 connected to the CODEC 411 or the SCSI controller 413. The NCU 415 exchanges information according to a predetermined procedure with an exchange that is directly connected to a telephone line and installed in a telephone office or the like.

Here, an example of fax transmission will be shown. The image signal for fax transmission is input from the reader unit 1 and the image memory unit 8 via the core unit 10.
The image signal transferred from the core unit 10 is input from the connector 400 and is output as a signal 453 by the memory controller 40.
Reaches four. The signal 453 is stored in the memory A 405 by the memory controller 404. The timing to be stored in the memory A 405 is generated by the timing generation circuit 409 according to the timing signal 459 from the reader unit 1. The CPU 412 connects the memory A 405 and the memory B 406 of the memory controller 404 to the bus line 463 of the CODEC 411. CODEC 411 is memory A
The image information is read from 405, encoded by the MR method, and the encoded information is written in the memory B406.
When the CODEC 411 encodes A4 size image information, the CPU 412 connects the memory B 406 of the memory controller 404 to the CPU bus 462. CPU4
12 sequentially reads the coded information from the memory B 406 and transfers it to the MODEM 414. And MODEM
414 modulates the encoded information and sends the fax information over the telephone line via the NCU.

Next, an example of fax reception will be described. The information sent from the telephone line is input to the NCU 415 via the connector 416 and is connected to the fax unit 4 by the NUC 415 in a predetermined procedure. The information from NCU 415 is demodulated by MODEM 414. CPU4
12 stores information from the MODEM 414 in the memory C407 via the CPU bus 462. When the information of one screen is stored in the memory C407, the CPU 412 controls the memory controller 404 to connect the data line 457 of the memory C407 to the line 463 of the CODEC 411. The CODEC 411 sequentially reads the coded information in the memory C407 and decodes it, that is, stores it in the memory D408 as image information. The CPU 412 communicates with the CPU 1003 of the core unit 10 via the dual port memory 410, and makes settings for outputting an image from the memory D 408 to the printer unit 2 or the image memory unit 8 via the core unit 10. Here, when the setting is completed, CP
The U 412 activates the timing generation circuit 409 and outputs a predetermined timing signal 460 to the memory controller 404.
Output to The memory controller 404 reads the image information from the memory D408 in synchronization with the signal from the timing generation circuit 409 and transmits it as a signal 452,
Output through the connector 400. The connector 400
Since the process from the output to the printer unit 2 or the image memory unit 8 has already been described in the core unit 10, the description thereof will be omitted.

<Explanation of File Unit 5> FIG. 6 is a block diagram showing the detailed structure of the file unit 5. The configuration and operation will be described with reference to FIG.

The file section 5 is connected to the core section 10 by the connector 500 and exchanges various signals. The image input signal 551 is input to the selector 521. Here, if the image input signal 551 is a multi-valued image that has not been subjected to compression processing, the selector 5 is determined by the CPU 516.
21 is switched and input to the compression circuit 503. The multi-valued image signal input to the compression circuit 503 is converted into compression information and output as a signal 552 to the memory controller 510. The signal 552 is a memory A 506, a memory B 507, a memory C 5 under the control of the memory controller 510.
08, the memory D509, or a combination of two sets of memories connected in cascade. On the other hand, when the signal 558 is selected by the selector 521, the compression processing is not performed and the signal is output to the memory controller 510.

Here, the memory controller 510 is C
A mode in which data is exchanged with the memory A 506, the memory B 507, the memory C 508, and the memory D 509 via the CPU bus 560 according to an instruction from the PU 516, and encoding / decoding
CODEC bus 570 of CODEC 517 for decoding
Mode for exchanging data via the memory and the memory A5
06, memory B507, memory C508, memory D50
9 is stored in one of the memories A506 to D509 under the control of the timing generation circuit 514 and the mode in which data is exchanged from the scaling circuit 511 via the bus 562 under the control of the DMA controller 518. It has five functions: a mode and a mode in which the memory contents are read from any one of the memories A506 to D509 and output to the signal line 558. In addition, memory A5
06, memory B507, memory C508, memory D50
Each of the 9 has a capacity of 2 Mbytes and 400 dpi.
An image corresponding to A4 is stored with the resolution of.

The timing generation circuit 514 is connected to the connector 5
00 and is connected to the control signal (H
SYNC, HEN, VSYNC, VEN) 553 to generate signals for accomplishing the following two functions. The first is to store information from the core unit 10 in the memory A50.
6 to a function of storing in any one of the memories D509 or two memories, and the second one is a memory A506 to
It has a function of reading image information from any one of the memories 509 and transmitting it as a signal 556. The dual port memory 515 is connected to the CPU 1003 of the core unit 10 and, at the same time, is a CP of the file unit 5 via the CPU bus 560.
It is connected to U516. Each CPU exchanges commands via the dual port memory 515. The SCSI controller 519 also interfaces with the external storage device 6 connected to the file unit 5 in FIG. The external storage device 6 is specifically composed of a magneto-optical disk, and stores data such as image information. The CODEC 517 is a memory A506 to a memory D5.
The image information stored in any one of the memory A09 and the memory 506 is read out and encoded by a desired MH, MR, or MMR method, and then stored in any one of the memory A506 to the memory D509 as encoded information. Also, it reads out the encoded information stored in the memories A506 to D509 and reads MH,
After decoding by the desired method of MR, MMR method,
The decoded information, that is, image information is stored in any of the memories A506 to D509.

An example of storing file information in the external storage device 6 will be described. The image signal from the core unit 10 is
The signal is input from the connector 500 and input to the selector 521 as a signal 551. Here, when the image input signal 551 is a multi-valued image that has not been subjected to compression processing, CP
The selector 521 is switched according to the determination of U516 and is input to the compression circuit 503. The multi-level signal input to the compression circuit 503 is converted into compression information and input to the memory controller 510. On the other hand, the signal 55
When 8 is selected, the compression processing is not performed and the data is input to the memory controller 510. The memory controller 510 uses the timing signal 559 generated by the timing generation circuit 559 according to the signal 553 from the core unit 10.
The compressed signal 552 is stored in the memory A 506 according to
The CPU 516 is the memory A5 of the memory controller 510.
06 and the memory B 507 are connected to the bus line 570 of the CODEC 517. CODEC 517 is a memory A5
The compressed information is read from 06 and encoded by the MR method, and the encoded information is written in the memory B507. C
When the ODEC 517 finishes encoding, the CPU 516 sets the memory B 507 of the memory controller 510 to the CPU 5
Connect to 60. The CPU 516 sequentially reads the encoded information from the memory B507 and transfers it to the SCSI controller 519. The SCSI controller 519 stores the encoded information 572 in the external storage device 6.

Next, an example of a case where information is taken out from the external storage device 6 and output to the printer unit 2 and the image memory unit 8 will be described. Upon receiving the information search / print command, the CPU 516 receives the encoded information from the external storage device 6 via the SCSI controller 519 and transfers the encoded information to the memory C508. At this time, the memory controller 510 sets the CPU bus 560 to the bus 566 of the memory C508 according to an instruction from the CPU 516.
Connect to When the transfer of the encoded information to the memory C508 is completed, the CPU 516 controls the memory controller 510 to connect the memory C508 and the memory D509 to the bus 570 of the CODEC 517.
The CODEC 517 reads the encoded information from the memory C508, sequentially decodes the encoded information, and then transfers the decoded information to the memory D509.

When magnification such as enlargement / reduction is required when outputting to the printer unit 2 and the image memory unit 8, the memory D509 is connected to the bus 562 of the scaling circuit 511, and the memory is controlled under the control of the DMA controller 518. D5
The contents of 09 are scaled. The CPU 516 communicates with the CPU 1003 of the core unit 10 via the dual port memory 515, and makes settings for printing an image from the memory D509 to the print unit 2 via the core unit 10. When the setting is completed, the CPU 516 activates the timing generation circuit 514, and the predetermined timing signal 559 is output.
Is output to the memory controller 510. The memory controller 510 reads the decoded information from the memory D509 in synchronization with the signal 559 from the timing generation circuit 514 and transmits it as a signal 556. The signal 556 is input to the selector 520, and it is determined whether or not the compression processing has been performed by the compression circuit 503 when being stored as file information in the file unit 5. This judgment is made by the CPU 516.
If the image information is compressed by the compression circuit 503, the image information is expanded by the expansion circuit 504 and then output as a signal 555. If it is not compressed by the compression circuit 503, the signal 557 is output as the signal 555. Then, the signal 555 is output to the core unit 10 via the connector 500. In addition, from the connector 500 to the printer unit 2 and the image memory unit 8
Since it has already been described in the core unit 10 until it is output to (2), it will be omitted.

<Description of LAN Interface Unit 7> FIG. 7 is a block diagram showing the configuration of the LAN interface unit 7. LAN interface section 7
The configuration and operation of will be described.

The LAN interface section 7 is connected to the core section 10 by the connector 700 and exchanges various signals. The input signal 754 is stored in the memory 704 under the control of the memory controller 705.

The memory controller 705 is the CPU 70.
In the mode in which the signal 754 is stored in the memory 704 under the control of the timing generation circuit 702 according to the instruction of No.
04, a mode for reading the memory contents and outputting as a signal 755, a mode for outputting the data stored in the memory 704 as a bidirectional signal 757 to the LAN I-F 707, and a mode for storing the bidirectional signal 760 in the memory 704. It has four functions.

The memory 704 asynchronously outputs the data input as the signal 754 to the LAN, or L
It is an intermediate buffer memory for asynchronously outputting the data input from the AN as a signal 755. The timing generation circuit 702 is connected to the connector 700,
Control signals (HSYNC, HEN, VS) from the core unit 10
YNC, VEN) to generate signals to achieve the following two functions. The first is the core portion 10
Is a function of storing information from the memory 704, and the second is a signal 755 for reading image information from the memory 704.
Is a function of transmitting as. Dual port memory 70
3 is connected to the CPU 1003 of the core unit 10 and the CPU 706 of the LAN interface unit 7. Each CPU exchanges commands via the dual port memory 703.

The LAN I-F707 is a connector B.
It has a control function for exchanging data on the LAN via 708, and controls data output to the LAN and data input from the LAN.

An example of transmitting image information input from the core unit 10 on the LAN in the LAN interface unit 7 will be described. Here, the image information input from the core unit 10 is transferred from either the reader unit 1 or the image memory unit 8. First, the CPU 706
Controls the memory controller 705 by signal 758 to connect the CPU bus to the bus of memory 704. The image signal of the core unit 10 transferred from either the reader unit 1 or the image memory unit 8 is input from the connector 700,
The signal 754 is input to the memory controller 705. The timing generation circuit 702 generates a timing signal 756 by the signal 752 from the core unit 10, and the memory controller 705 temporarily stores the signal 754 in the memory 704 in accordance with this signal. Subsequently, the CPU 706 controls the memory controller 705 to connect the bidirectional bus to the memory 704.
Connect to the bus. The image information temporarily stored in the memory 704 is sent as a bidirectional signal 757 to the LAN I-F 707, and further output as a bidirectional signal 759 from the connector B 708 onto the LAN by signal control in the LAN I-F 707.

Next, an example in which the LAN interface section 7 outputs the image information input from the LAN to the core section 10 will be described. Here, the image information output to the core unit 10 is transferred to either the printer unit 2 or the image memory unit 8. First, the CPU 706 controls the memory controller 705 by the signal 758 to connect the data line of the memory 704 to the bidirectional bus. Connector B70
Image information input from LAN from 8 is a bidirectional signal 75
9 is sent to the LAN I-F707, and the LAN
It is temporarily stored in the memory 704 as the bidirectional signal 757 by the signal control by the I-F 707. Then, CPU7
The memory controller 705 controls the memory controller 705.
Connect the data line of. The CPU 706 is the CPU 1003 of the core unit 10 via the dual port memory 703.
And the setting for outputting the image information from the memory 704 through the core unit 10 to either the printer unit 2 or the image memory unit 8 is performed. When setting is completed, CP
The U 706 activates the timing generation circuit 702 and outputs a predetermined timing signal as a signal 756 to the memory controller 705. The memory controller 705 is
The image information is read from the memory 704 in synchronization with the signal from the timing generation circuit 702 and output to the connector 700 as a signal 755. The output from the connector 700 to the printer unit 2 and the image memory unit 8 has already been described in the core unit 10 and will not be described.

<Description of Image Memory Unit 8> FIG. 8 is a block diagram showing the configuration of the image memory unit 8. The configuration and operation of the image memory unit 8 will be described with reference to this figure.

The image memory unit 8 is connected to the core unit 10 by the connector 800 and exchanges various signals. The input signal 854 is stored in the memory 804 under the control of the memory controller 805.

The memory controller 805 is the CPU 80.
6, a mode for exchanging data with the memory 804 and the CPU bus 857, and a timing generation circuit 8
The mode in which the signal 854 is stored in the memory 804 under the control of 02 and the memory content is read out from the memory 804.
It has three functions of a mode of outputting as 55.

The memory 804 has a resolution of 400 dpi,
In addition to the memory capacity of 32 Mbytes capable of storing an image corresponding to A3 with 256 gradations, various image processing, binarization processing, image compression processing, block selection processing, etc. described later can be executed. It also has a sufficient work memory.

The timing generation circuit 802 is connected to the connector 8
00 and is connected to the control signal (H
SYNC, HEN, VSYNC, VEN) to generate signals for performing the following two functions.
The first is a function of storing information from the core unit 10 in the memory 804, and the second is a function of transmitting image information from the memory 804 as a read signal 855. The dual port memory 803 is the CPU 1003 of the core unit 10.
Is connected to the CPU 806 of the image memory unit 8. Each CPU exchanges commands via the dual port memory 803.

In the image memory unit 8, the core unit 10
An example of accumulating more input image information will be described. The image information input from the core unit 10 is transferred from any one of the reader unit 1, the fax unit 4, the file unit 5, and the LAN interface unit 7. First C
The PU 806 controls the memory controller 805 to control the CPU.
The bus is connected to the bus of the memory 804. The image signal of the core unit 10 transferred from any one of the reader unit 1, the fax unit 4, the file unit 5, and the LAN interface unit 7,
The signal is input from the connector 800 and is input to the memory controller 805 as a signal 854. The timing generation circuit 802 generates a timing signal 856 from the signal 852 from the core unit 10, and the memory controller 805 stores the signal 854 in the memory 804 in accordance with this signal.

Next, the memory 804 of the image memory unit 8
An example of outputting the image information stored in the core unit 10 to the core unit 10 will be described. Here, the image information output to the core unit 10 is
The data is transferred to any of the printer unit 2, the fax unit 4, the file unit 5, and the LAN interface unit 7. First CP
U806 controls the memory controller 805 to connect the data lines of the memory 804. The CPU 806 is a CPU of the core unit 10 via the dual port memory 803.
1003 communicates with the memory unit 804, passes through the core unit 10, and passes through the printer unit 2, the fax unit 4, the file unit 5, and the L unit.
Settings are made to output image information to any of the AN interface units 7. When setting is completed, CPU80
6 activates the timing generation circuit 802 and outputs a predetermined timing signal from the signal line 856 to the memory controller 805. The memory controller 805 synchronizes with the memory 80 in synchronization with the signal from the timing generation circuit 802.
The image information is read from No. 4, transmitted as a signal 855, and output to the connector 800. Since the output from the connector 800 to the printer unit 2, the fax unit 4, the file unit 5, and the LAN interface unit 7 has already been described in the core unit 10, the description thereof will be omitted.

Further, the CPU 806 of the image memory unit 8
Has a function of performing image binarization processing on a partial area in the multi-valued image stored in the memory 804. CPU
As a method of image binarization processing by 806, a method by simple binarization is used. The binarized image obtained as a result can be stored in the memory 804 again.

Further, the CPU 806 of the image memory unit 8
Is an attribute of a partial area in the image stored in the memory 804, that is, one of a character part, a title part, a frame part, a front part, a halftone figure part, a line figure part, and a line part. It has a function to perform block selection processing, which is a classification processing.

An example of the block selection process in the image memory unit 8 will be described in detail below.

First, the image data to be subjected to the block selection processing is stored in the memory 804. The image data in this case may be read by the reader unit 1 or may be image data transferred from the fax unit 4, the file unit 5, and the LAN interface unit 7. However, since the image data that can be subjected to the block selection processing is limited to binary image data, when the image data is read by the reader unit 1, the fax unit 4, the file unit 5, and the LAN interface unit. If the image data transferred from 7 is a multi-valued image, it is necessary to binarize the image data via the binarization circuit 1012 and then store it in the memory 804. When performing block selection processing on multi-valued image data stored in the memory 804 of the image memory unit 8 in advance, the binarization processing function of the CPU 806 is used to binarize the image data. Must be stored in the memory 804.

In this way, the two stored in the memory 804 are stored.
Block selection processing is performed on the value image data. FIG. 9 is a schematic flowchart showing an algorithm of the block selection process.

First, in step S900, if it is desired to increase the processing speed of block selection, image data is thinned out. If the image data is thinned out, the block selection processing is performed on the thinned image.

Concrete image data thinning is performed by examining the connectivity of black pixels in an m × m block of pixels. For example, if there are two connected black pixels in a 3 × 3 pixel block, the 3 × 3 pixel
The pixel block of is thinned out to one black pixel. On the contrary, when two white pixels are connected in the block, the block is thinned out to one white pixel.

Next, in step S901, the image of the pixels is analyzed to search for connectivity, and the pixels are classified according to their size and relative position to other connected components. A connected component is a set of black pixels completely surrounded by white pixels. Therefore, one black pixel connected component is completely separated from other black pixel connected components by at least one white pixel. This step S90
The details of No. 1 will be described with reference to FIG. 10. In general, a class of connected components that is performed based on a search for connected components and size information and some statistical information obtained from the connected components. Divide. In this classification, each connected component is first classified into a text unit or a non-text unit. And the non-text unit is
Further detailed analysis is performed to determine whether the data has a frame structure, a halftone image, a line drawing, a table, or other tabular text data. As a result, if unknown, it is not classified as unknown. A hierarchical tree structure is then created for each connected component to provide the connected component's constituent data and facilitate reconstruction of that data.

Next, in step S902, adjacent connected components are grouped unless a gap line is sandwiched therebetween. The grouping here may be performed in the vertical direction or in the horizontal direction. This corresponds to whether the text units to be grouped are vertical writing or horizontal writing.The distance between connected components that are close to each other in both directions is checked in advance in the horizontal and vertical directions. Grouping is performed in the horizontal direction when the distance in the direction is small, and in the vertical direction when the distance in the vertical direction is small. The tree structure generated in step S901 described above is used to prevent improper mixing of text and non-text. Further, in step S902, it is determined whether the text units are grouped into lines by detecting vertically or horizontally extending gaps between lines and vertically extending boundaries of non-text units. To do. This row structure is maintained in the tree structure by appropriately updating the hierarchical tree structure.

Next, in step S903, if there is a small space in the direction opposite to the direction in which the rows grouped in step S902 above are grouped,
It is grouped again in that direction and blocked. On the other hand, the non-text unit is used as a boundary for the image page. Text units between two non-text units are processed separately from other text line units.

Further, in step S903, an analysis is performed to determine whether the non-text unit that could not be classified in step S901 is a title having a large font size. If they are determined to be titles, those units are given the appropriate attributes and the tree is updated. The title helps restructure the page.

10 to 12 show the above-mentioned step S9.
21 is a flowchart showing the processing of No. 01, showing how connected pixels of pixel image data are detected and these connected pixels are classified.

In step S1001, the pixel image data is searched for by contour line tracking. Contour tracking is done by scanning the image, as shown in FIG. The scanning starts from the lower right as indicated by arrow A and proceeds upward until the right edge of the figure is encountered. Note that this scan may be performed in another direction, for example, from the upper left to the lower right. When it hits a black pixel, whether or not the adjacent pixel is a black pixel is checked in order of the direction of the pattern indicated by 31 in FIG. This search is called an eight-direction search because it is represented by a vector in eight directions when viewed from the center. If there are adjacent black pixels, this process results in the outer contour of the figure. Specifically, as shown in FIG. 14, the scan in the direction of arrow A hits the point corresponding to the end of the letter "Q" 32. Therefore, the investigation of the adjacent pixels is performed by the above-mentioned pattern 31, and the outer contour of the character "Q" is traced. The inside of the closed contour is not tracked.

Next, the contour line obtained by the eight-direction search,
That is, when one connected component is extracted, scanning is continued until the next black pixel is encountered. Then, for example, the object 34 that seems to represent a complete black area is searched in eight directions. Similarly, a non-text object 35, which is the handwritten character "non-text", is tracked, and then a collection object 36 of this character that forms the word "text" is tracked. The scan shown in FIG. 13 is continued until all connected components are detected.

Then, in step S1002, all connected components are cut out into rectangles. In this case, the smallest possible rectangle covering the individual connected components will be drawn. Thus, the rectangle 37 around the object 32 and the rectangle 39 around the object 34 in FIG.
A rectangle 40 is drawn around the object 35. Also,
Text objects 36a, 36b, 36c, 36d
The same applies to rectangles 41a to 41d for.

Next, in step S1003, all the rectangles are positioned in the tree structure. In most cases, the tree structure obtained in step S1003 results directly from the root for each object. This is because only the outer contour of the connected component is tracked, not the interior of the closed region. As a result, the rectangle 37 corresponding to the connected component 32 as shown in FIG. 15 is directly generated from the root of the page. However, the rectangle 4 surrounding the non-text object 35
Rectangle 4 surrounding 0 and text objects 36a and 36b
When the rectangle is completely contained in another rectangle like 1a and 41b, these connected components are children of the included connected component (in this case, the component 34). In addition, ingredient 3
Each connected component that has at least one child, such as 4, makes the component itself a "primary child." In the example of FIG. 15, the component 39 includes itself as a main child together with the other child components 40, 41a, 41b.

Next, in step S1004, each first level connected component in the tree is classified into a text unit or a non-text unit. This classification process consists of two steps. In the first step, the rectangle surrounding the connected component is compared to a predetermined size. Here, when the height of the rectangle surrounding the connected component exceeds a predetermined value corresponding to the maximum value of the font size, or when the width of the rectangle is the empirically determined constant value (“5”).
, The connected component is classified as a non-text unit, and
Give the unit a "non-text" attribute.

In the second step, all remaining un-attributed units, ie those not classified as non-text, are assigned a value determined based on the statistical size obtained from all remaining connected components. Compare. In particular, it calculates the average height of all rectangles that were not considered non-text. An adaptive threshold is obtained by multiplying this average height by a certain value (generally "2"). All units above this threshold are classified as non-text. On the other hand, units smaller than this threshold are regarded as text. In this way, each unit is classified and given the appropriate attributes. Then, some processing is further performed following the above-mentioned two classifications. This is described in more detail below.

As described above, after all units at the first level of the tree have been classified as text or non-text, the text unit's children are classified as text, including the major children (ie, themselves). To be done. The main non-text children are left as non-text, while the other children are classified as text.

Next, in step S1005, the first unit is selected, and if the unit is text in step S1006, the process advances to step S1007 to select the next unit. Then, the processes of steps S1006 to S1007 are repeated until the non-text unit is selected, and when the non-text unit is selected, the step S1 is performed.
Proceed to 008.

In step S1008, it is checked whether the non-text unit has a child. For example,
In the example shown in FIG. 15, the non-text unit 39 is a non-text main child 39 and texts 40 and 41.
I have children a and 41b. If there is a child in that unit, the process advances to step S1009,
Filters whether the unit is halftone (or grayscale). This halftone filtering looks at its children and counts those that are smaller than the "noise" size. Here, the unit of "noise size" is one whose height is smaller than the minimum font size in the input image data. If the number of children smaller than the noise size is greater than half the total number of children, the unit is determined to be a halftone image. As a result, the process advances from step S1010 to S1011 to give an attribute of “halftone”, and then to step S1.
At 012, examine the text in the halftone image. That is, among the children of the halftone image, those of text size are not children of the halftone image, and the tree structure is modified so as to be at the same level as the halftone image. If this procedure is appropriate, character recognition in the halftone image is also possible. Then, step S10
Returning to 07, the next unit is selected and the above processing is repeated.

If the unit is not halftone as a result of the halftone filtering in step S1009, the process advances from step S1010 to S1013 to select the main child of the unit for further processing. Then, the process proceeds to step S1014.

On the other hand, if it is determined in the above step S1008 that it is a non-text unit and has no child, or if the main child is selected for the subsequent processing in step S1013, that step is executed in step S1014. Perform frame filtering for units. This frame filtering is a process of determining whether or not the unit is a frame. Here, being a frame means that there are parallel straight lines having a width and a height of almost the same length that form a rectangle surrounding the unit. In particular, in the target unit, when the non-text unit 42 as shown in FIG. 16 includes the connected component 43 having the contour component as shown by 44, the line width of this connected component in the row i is x, that is, Left edge of the contour line 45
The distance is from a to the right end 45b. Also in line j,
There are two line widths inside the connected component. That is, 46
Between a-46b and between 47a-47b. The distance between 47a and 47b, which is the longest line width, is defined as the distance x.

The distance x is calculated for every line in the non-text unit 42 and the following inequality is used to check whether the non-text unit is a frame.

Σ (Xk−W) ** 2 / N <[threshold value]; k = 1 to N where Xk is the longest line width in the kth row in the connected component, W, as described above. Is the width of the rectangle 42, N is the number of rows, and the threshold value is a value calculated in advance so that it can be detected as a frame even if the frame is slightly inclined. Also, to allow a tilt of 1 °, sin (1 °) × L + [constant value]
The threshold value of This constant value means step S1.
The average height of the characters calculated in 104.

If the above inequality is satisfied, the unit is determined to be frame data, and step S1015 is performed.
From S1016, the attribute "frame" is added. This allows a determination such as "frame and table" or "frame and halftone" for the frame.

When the processing of step S1016 is completed,
It is checked whether or not the frame data includes a table or a table format. That is, in step S1017, an inspection is performed to obtain the white contour in the connected component. The white contour is basically the same as the (black) contour obtained in step S1001, but it is obtained by checking white pixels instead of black pixels. As shown in FIG. 18, the inside of the non-text unit is searched in the direction of arrow B from the lower right to the upper left. When a white pixel is first encountered, an outward search such as 51 is performed from that point on the neighboring white pixels. At this time, it should be noted that in the outward search of 51, only the directions from 1 to 4 are necessary. That is, the white contour tracking process here is a four-direction search. This process continues until all white contours have been retrieved. For example, white contour tracking is performed by black line segments 52, 53,
The outline portion surrounded by 54 and 55 is taken out, and the same processing is performed on the inside of a black pixel such as 56. The above-described scan in the direction of arrow B is continued until all closed white contours in the non-text object are tracked.

Next, in step S1018, the density of non-text units is calculated. This density is calculated by counting the number of black pixels in the connected component and dividing the number of black pixels by the total number of pixels surrounded by the rectangle. Then, in step S1019, the number of white contours in the searched non-text unit is determined. As a result, if the number of white contours is 4 or more, it is determined that the non-text image is a table or text blocks arranged in a table shape, and the filling rate of white contours is calculated in step S1020. The filling rate of the white outline indicates the ratio of the area surrounded by the white outline in the non-text image. In the example shown in FIG. 18, there are white contours such as 57 and 59 that are completely composed of white pixels, and 6
There is also a white outline including a black pixel area such as 0 and 61. Here, if the fill factor is high, the non-text image is probably a table or text blocks arranged in a table. In order to make this estimation more reliable, it is examined whether or not the target white contour has a lattice-like internal structure in the horizontal and vertical directions. In particular, step S
A white contour having a boundary line that does not traverse at least two contour lines in the horizontal or vertical direction at 1022 is regarded as not in a grid shape and is recombined. For example, FIG.
In the example shown in (1), the left boundary 62 and the right boundary 63 of the white outline 59 extend in the vertical direction so as to coincide with the left boundary 64 and the right boundary 65 of another white pixel 60. Therefore, these white contours are judged to be arranged in a grid and are not recombined. Similarly, the upper boundary 66 and the lower boundary 67 of the white contour 63
Extends in the horizontal direction so as to coincide with the upper boundary 68 and the lower boundary 69 of another white pixel 70. As a result, it is determined that these white contours are also arranged in a lattice, and are not recombined.

19 to 21 are diagrams for explaining a case where white contours are combined. FIG. 19 shows an example of a non-text unit in which the non-text unit 71 includes units from a halftone image to a binary image. In the figure, the non-text image 71 includes a black pixel area 72 and white pixel areas 74 to 79. That is,
In the example of FIG. 19, since the filling rate of the white pixel area is sufficiently high, the white pixel area is recombined in step S1021 and the process proceeds to step S1022. First, the upper and lower ends of the white contour 74 are compared with the upper and lower ends of the white contour 77. In this example, their upper and lower edges do not coincide, so 75 and 76 are joined to create a new white contour 76 ', as shown in FIG.

Next, the left and right boundaries of the white contour 77 are set to the white contour 7.
Compare with the left and right boundaries of 8. Since these boundaries do not match, 77 and 79 are recombined into a new white contour 77 ', as shown in FIG. This process is then repeated horizontally and vertically until recombination no longer occurs.

As described above, the white contours of the table are difficult to combine, and things other than the table, such as halftone images and line graphics, are easy to combine. Next, in step S1023, the recombination rate is calculated. As a result, if the recombination rate is high or the number of white contours remaining after the recombination process is less than 4, the process proceeds to step S1028, and the non-text unit is determined to be a halftone image or a line figure. This processing will be further described later.

If the recombination rate is not high or four or more white contours remain in step S1023, the process proceeds to step S1024 and a table is determined. Next, in step S1025, the inside of the newly determined table is examined to search or classify the connected components included therein.
Then, in step S1026, the tree structure is updated according to the new internally connected component. Next, in step S1027, the internally connected component is classified again as text or non-text,
Add the appropriate attributes. This processing is the same as the processing of steps S1002 to S1004 described above. Then, the process returns to step S1007 to select the next text unit.

Here, steps S1021 and S10 are executed again.
Returning to 23, if the filling rate is not high or the recombination rate is not high, the non-text unit (here, a frame graphic) is likely to be a halftone image or a line graphic.
Therefore, in step S1028, it is determined whether the unit is a halftone image or a line figure based on the average run length of black pixels in the unit in the horizontal direction, the ratio of white pixels to black pixels, and the density. Generally, a very dark image is considered to be a halftone image, and a white and bright image is determined to be a line figure. Particularly, when the average run length of the white pixels is almost “0” (almost black or a mottled image), and when the density obtained in step S1018 is larger in black than in white (that is, the density is If the threshold value is greater than about 0.5 (which is the first threshold value), the frame unit is determined to be a halftone. If the density is not higher than the first threshold value, the unit is determined to be a line figure.

If the average run length of the white pixels is almost "0" and the average run length of the white pixels is larger than the average run length of the black pixels, the frame unit is determined to be a line figure. . However, when the average run length of white pixels is larger than the average run length of black pixels, the frame unit is determined to be a line figure.
Also, if the average run length of white pixels is not greater than the average run length of black pixels (ie, black is also the dominant image), then more detailed testing is required.

In particular, when the number of black pixels is much smaller than the number of white pixels (that is, the number of black pixels is less than twice the number of white pixels (let's call this the second threshold value)), this frame The unit is determined to be halftone. On the other hand, if the value obtained by dividing the number of black pixels by the number of white pixels is not greater than the second threshold value, but the density calculated in step S1018 is greater than the first threshold value, the frame unit is Judge as a toned image. Otherwise,
Judge as a line figure.

As described above, when the frame unit is determined to be a line figure in step S1028, step S102.
9, the attribute "line figure" is added, and all the children are removed in step S1030. In particular, once a unit is determined to be a line figure, character recognition processing is not performed again for that unit. After that, the process is returned to the above step S1007, and the next text unit is selected.

On the other hand, if it is determined in step S1028 that the frame unit is not a line figure, step S1031.
Go to, add the attribute "halftone" here,
In step S1032, the child of the text size is removed from the children of the unit. Note that all children larger than the text size are allowed to remain children of the frame halftone image. Then, the above-described step S
The processing is returned to 1007, and the next text unit is selected.

Here, returning to step S1019, if the number of white contour components is not larger than 4, it is determined that the frame unit is not a table. Then, the process proceeds to step 1033, and the density obtained in step S1018 is compared with a certain threshold value (about 0.5). This threshold is chosen because the text units and line graphics in a frame should be less than half of all pixels. If the density is smaller than this threshold, step S103
Proceed to step 4 to examine the internal structure of the frame unit. This processing is similar to the processing for the internal structure of the frame unit in step S1001 already described.

If it is determined in step S1033 that the density is not smaller than the predetermined threshold value, step S1.
Proceed to 042 to determine whether the frame unit is classified as a line graphic or a halftone image, or the frame cannot be classified (that is, the frame is “unknown”).
Is determined.

Returning to step S1015, step S1
If no frame is detected in the non-text unit by the frame filtering in 014, the process proceeds to step S1035, and a "line" is added to the non-text unit.
It is determined whether or not is included. A "line" is a non-text unit that is useful for demarcating text. However, the text delimited (enclosed) by such a line is often very close to that line, and contact may have occurred. Therefore, it is necessary to consider both the case where the text is in contact and the case where the text is not in contact as the line detection processing.

First, in order to search for a line when no contact has occurred, a vertical histogram of a non-text unit is obtained. In the case of (A) shown in FIG. 17, the histogram 48 of the line should have a uniform value whose height is almost equal to the line width. The line width is approximately equal to the width of the text unit (“W”), but if there is a difference, it is due to the slope θs. The inclination is set when the original is input. Therefore, the height 49 of each cell k in the histogram is compared with the width W to determine whether the non-text unit contains a line. Specifically, it is a process of comparing the mean square of the difference between these values with a threshold value as in the following equation.

Σ (cell k−W) ** 2 / N <[threshold value]; k = 1 to N This threshold value is calculated so as to allow the twist or the inclination θs of the line in the non-text. It It has been found that a threshold of Σ [k · sin (1 °) / N] ** 2 gives a satisfactory result for a twist or inclination of 1 °.

When the non-contact line cannot be detected by the above inequality, a search is made as to whether or not the line in contact is included. In order to check whether or not the touching line is included in the non-text unit of interest, it is necessary to check whether or not a linear object exists near the boundary line of the unit. For example, in the case of (B) shown in FIG. 17, a line exists near the boundary line of the rectangle surrounding the unit.
In this case, it can be checked by calculating the sum of squares of the distance from the boundary line. That is, in this case, the following inequality is calculated.

Σ (Xk) ** 2 / N <[threshold value]; k = 1 to N Here, if the left side is smaller than a predetermined threshold value, there is a line in contact. Recognize. As this threshold value, the same value as in the case of a line that does not make contact may be used.

If a line is detected in step S1035,
The process advances from step S1036 to step S1049 to add the attribute "line" to the non-text unit. And
Returning to step S1007 described above, the next unit is selected. If no line is detected in step S1035, the flow advances from step S1036 to S1037 to check the size of the non-text unit. Here, if the size is smaller than a certain threshold value, the classification of the non-text unit cannot be determined. Therefore, the processing proceeds to step S1038, and the attribute of "unknown" is added.
The threshold is determined from the maximum font size. Specifically, a good result can be obtained by setting the value to half the maximum font size. Then, the process returns to step S1007 to select the next unit.

If it is determined in step S1037 that the size is larger than the predetermined threshold value, the flow advances to step S1039 to search for a white contour in the internal area of the non-text unit, and steps S1017 to S10.
The number of white contours is calculated as described in 19.

Then, in step S1041, if the number of white contours is not 4 or more, the process proceeds to step S1042, and the size is calculated in order to confirm whether or not the size is sufficient as a line figure or a halftone image. The size is determined based on the height, width, and maximum run length of black pixels of the text unit. In particular, if the height and width of the non-text unit is smaller than the maximum font size, it is considered that the unit is not large enough to form a line drawing or a halftone image, and the process proceeds to step S1043, and "unknown" Add the attribute of. Similarly, when the unit width is larger than the maximum font size but the maximum length of the black pixel run length is not larger than the maximum font size, the process also proceeds to step S1043 and the attribute of "unknown" is added. Then, the process returns to step S1007,
Select a new unit.

If it is determined in step S1042 that the non-text unit is large enough to form a line figure or a halftone image, the flow advances to step S1044 to add an attribute of a line figure or a halftone image. Then, steps S1044 to S1048 are the same as steps S1028 to S1032, and a description thereof will be omitted.

The flow described above (step S90 in FIG. 9)
All the connected components in the input image are examined according to the process 1), and when the classification is completed, a tree structure as shown in FIG. 28 is obtained. As shown, the route corresponds to the page of the input image. The root children are text blocks or non-text blocks consisting of unknowns, frames, pictures (figure) and lines. The children of the frame are text blocks, "unknown" non-text data, tables containing text blocks,
Photo (figure), line.

FIG. 24 is a diagram showing a page 90 of pixel image data. In this, a large font size text 91, for example, a table 92 including text data as shown in 93, text data 94, a horizontal line 95,
Line title 9 with another title 96, text data 97 consisting of two paragraphs, caption 99 and surrounded by a frame
8, a second column that starts with a title 100 and continues to the text data 101, a halftone image 102 surrounded by a frame having a caption 103, and text data 10
4, the horizontal line 105, and the final paragraph 106 are included. FIG.
Indicates that the same image has been subjected to the processing in step S901 described above. As is clear from the figure, the connected component in the page 90 is cut out in a rectangular shape, and the attribute of the inside is checked by the processing shown in steps S1015 to S1034.

All the text units obtained in step S901 are grouped in the vertical direction or the horizontal direction in step S902 regardless of the position in the tree. This grouping operation is based on the cohesion of each text unit and its surrounding units. Also, gaps (spatial spaces) that appear to represent columns are detected and maintained in both vertical and horizontal directions. The detailed process of step S902 will be described below with reference to FIG.

First, in step S2201, the boundary line of the non-text unit is extended vertically and horizontally to form a gap line marker. As shown in FIG.
Vertical gap line markers 109a, 109b
Is stretched vertically until it intersects a text or non-text unit (unit 95 in this example).
Similarly, the gap line markers 109c, 109
d is also extended until it intersects unit 95. The same process is performed for the horizontal gap line marker. The gap line marker is effective for detecting a gap (spatial space), and thereby a column can be obtained.

Next, in step S2202, as shown in FIG.
As shown at 107 in No. 5, line joining of text units is performed. In this connection, the distance between connected components that are close to each other in both directions is checked in advance in the horizontal and vertical directions. If the horizontal distance is small, the horizontal direction is used. If the vertical distance is small, the vertical direction is used. Done. The combination direction corresponds to whether the text units to be combined are set vertically or horizontally.

Now, these text units are combined into one text line if the following conditions are met:

(1) The bond does not cross the gap line marker.

(2) The text unit is in contact with another text unit or at a distance equal to or less than a certain threshold value. This threshold value may be obtained by multiplying the average length of the text obtained in step S1004 by another scale factor experimentally obtained (a satisfactory result is obtained at 1.2).

However, before joining, the gap between the text units is extended in the horizontal direction when the text units are in horizontal writing mode, and in the vertical direction when the text units are in vertical writing mode, thereby extending in the direction showing the column structure. Can determine whether there is a gap. For example, in the example of FIG. 25, the gap 108 exists between two texts. The gap extends in the vertical direction over several lines, and therefore, in step S2202, even if the distance between the text units is less than or equal to the threshold value, the gap is left.

Next, in step S2203, the set of text units that were not combined in step S2202 are overlapped by other text units that are adjacent to them, and the combination does not cross the gap line marker. When the binding is done. This step is effective in erasing anything that does not come from the structure of paragraphs, but simply from the relationship of spaces in a text line. In the example of FIG. 25, the gap 108 left in step S2202 is erased in step S2203. This is because it overlaps the characters in the line immediately below and does not cross the gap line marker. Then, in step S2204, the tree structure is updated.

FIG. 26 is a schematic diagram showing the result of the grouping process described in step S902, and FIG.
FIG. 9 is a diagram showing how the tree structure is changed by the process of step S902. As shown in FIG.
Combined text units are grouped into 110
Is converted to a line of text like. In particular, text units are always combined into text lines anywhere in the tree structure. For example, 111 is below the frame table on the tree structure, but is also joined. However, it should be noted that the regrouping beyond the white contour obtained in steps S1017 to S1039 is not performed. This is to prevent the items in the table from being put into one line. The gap between the left and right columns is maintained. Also, non-text units are not regrouped. Therefore, they are not grouped even if they are at distances equal to or less than the threshold value like 112 and 113.

FIG. 29 shows a tree structure in which this new grouping is reflected.

After the text units are combined to form a text line in the process described in FIG. 22 (step S902 in FIG. 9), the text line is combined in the direction opposite to the combination direction when the text line is formed in step S903. It becomes a text block. This process will be described in more detail with reference to FIG. The process of grouping depends on the unity of text line units and the position of non-text units. For example, intervening non-text lines act as borders,
Prevents text lines on the opposite side from grouping together into one text block. All text lines between two consecutive non-text line units are processed simultaneously. Additionally, in step S903, some text units should be combined with non-text units (eg, text captions that are composed with non-text images), or some non-text units should be combined with other non-text units (eg, text units). , Which is the line shape associated with the halftone image),
Be examined.

FIG. 23 is a flowchart showing how text lines are grouped into text blocks. First, in step S2301, a title block is formed from those classified as non-text units in step S1004 described above. The criterion is that it is smaller than the maximum font size but larger than the average text size. Thus, a title block is formed by grouping all non-text units of similar size and proximity. Then, the attribute "title" is added to the block. The attribute of picture text is added to all the remaining non-text blocks that cannot be grouped. The tree structure is updated appropriately accordingly. The title is useful for page reorganization.

Next, in step S2302, non-text units between text lines are detected. These non-text units act as boundaries between text blocks, preventing text lines from becoming a single text block.

In step S2303, the text lines are grouped into a text block by a process consisting of two steps, in the direction opposite to the direction in which the text lines are formed (hereinafter referred to as "block combination direction"). . The first step is to search for gaps between columns. For that purpose, for example, a histogram in the block combination direction of pixels is calculated. In the next step, the distance between the text lines that are continuous in the block joining direction is calculated as step S1.
If the text height is smaller than that calculated in 004, these text lines are grouped within each column. Step S2303 is useful for joining text lines that follow the same paragraph, such as text line 114 in FIG.

Next, in step S2304, vertically or horizontally adjacent text blocks are not separated by non-text units, and
If it does not break any gaps found from the histogram obtained at 2303, it is grouped.

The grouping of the text blocks is performed based on the separation state between the blocks which is smaller than a certain threshold calculated according to the height in the vertical direction calculated in step S1004. In the example of FIG. 26, paragraph 115
Step S230 for grouping text blocks made up of text lines of paragraphs and text lines of paragraph 116.
4 is effective. However, it is not effective to combine 117 and 118. This is because these text blocks are separated by non-text blocks 119 (lines). Step S2305 determines whether one text block should be combined with a non-text block or one non-text block should be combined with another non-text block. Text blocks can be combined with non-text title blocks, non-text halftone blocks, and non-text lines with attachments. These combinations are performed according to the following judgment. (1-a) If a text block overlaps a non-text title block horizontally and vertically, the text block is combined with the non-text title block (provided that the text block and the text block Both title blocks are in horizontal writing mode). (1-b) If a text block is close to a non-text title block in the vertical direction and overlaps in the horizontal direction, the text block is combined with the non-text title block (however, the text block And the title block are both vertically set). (2) If a text block is smaller (horizontally or vertically) than a word-sized block and the text block has no adjacent word-sized text blocks, this text block is a non-text halftone. It is placed inside the image block. (3) For a text block that overlaps a non-text line with an appendage, the line with the appendage is simply text because it is probably underlined text.

Also, some non-text blocks are combined with other non-text blocks according to the combination shown in FIG.

In FIG. 35, Test is
The contents are as follows.

Test # 1: Combine if one block is completely contained in another block. Test # 2: Combine if the width of the picture text is smaller than the width of the word size block.

Test # 3: If blocks are close to each other, they are combined.

Then, in step S2306, the attribute is modified and the tree structure is updated by the processing described so far.

FIG. 27 shows the block structure obtained by the processing of FIG. 23, and FIG. 30 shows an example of the tree structure. As the blocks in FIG. 27, the title block 1
20, a text block 121, and a halftone graphic / line graphic 122. Further, as the frame line data, there are a table format 123 and a table structure 124 having a text unit 125 inside. The non-text image 127 serves as a separator for various units in FIG.

The block selection process is performed in the image memory unit 8 according to the above-described algorithm. By this block selection processing, each unit based on the black pixel connected component in the image
Frame part, front part, halftone figure part, line figure part, line part,
Block selection processing result information based on the attribute classification information identified by any of the above and area information indicating the coordinates and size of the smallest rectangle surrounding each unit on the image. The block selection processing result information is stored in the memory 804.
Will be recorded temporarily.

Further, the CPU 806 of the image memory unit 8
Has a function of performing character recognition processing on the image information data stored in the memory 804.

Next, an example of character recognition processing in the image memory unit 8 will be described in detail.

First, the document image data to be subjected to the character recognition processing is stored in the memory 804. The image data in this case may be read by the reader unit 1, or may be image data transferred from the fax unit 4, the file unit 5, or the LAN interface unit 7. However, since the image data that can be subjected to the block selection process is limited to binary image data, when the image data is read by the reader unit 1, the fax unit 4, the file unit 5, or the LAN interface unit. 7 is multi-valued image data, it is necessary to binarize the image data via the binarization circuit 1012 and then store it in the memory 804. Further, as the document image data to be subjected to the character recognition processing, it is possible to perform the character recognition processing only on the partial area which is determined to be the character portion or the title portion in the result of the block selection processing described above. Is.

Character recognition processing is performed on the image data thus stored in the memory 804. The algorithm of the character recognition processing according to the present embodiment will be described in detail with reference to the flowchart shown in FIG.

First, in step S3101, character cutting processing is performed on the image data. The character clipping process detects the image area of each character in the image data,
This is performed by obtaining the circumscribed rectangular image as a character cutout image. Here, in principle, the image of one character unit is composed of connected pixel components. However, when the connected pixel component is considerably smaller than the size of the adjacent connected pixel component, another connected pixel is used. It is combined with the component to form a single character image or treated as a noise image. On the other hand, if the connected pixel component is considerably larger than the size of the adjacent connected pixel component, a vertical or horizontal projection histogram is obtained, and if there is a portion with a small histogram value, that portion is taken as the boundary, Separation of one connected pixel component is performed. FIG. 32 shows an example of character cutting processing by separating the connected pixel components. The cut-out character image data obtained as described above is temporarily stored in the memory 804.

Next, in step S3102, normalization processing is performed on the cut-out character image data subjected to the character cutting processing in step S3101. In the normalization processing, first, the size of the cut-out character image is enlarged or reduced to a certain reference rectangular size. Next, the inclination, line width, density, etc. of the cut-out character image are corrected to a state in which matching with a standard pattern in a previously prepared recognition dictionary is most easily obtained.

Next, in step S3103, a recognition process is performed on the character image subjected to the normalization process in step S3102. In the recognition process, first, a feature vector is extracted based on a contour feature or the like. Then, the obtained feature vector and the standard pattern vector of all recognition target characters in the prepared recognition dictionary are matched, and the character code corresponding to the most probable standard pattern is used as the recognition result. It is stored in the memory 804. At this time, the degree of certainty of pattern matching in the character code as the recognition result (hereinafter, this is referred to as a pattern matching degree) is also stored.

Further, the CPU 806 of the image memory unit 8
The LAN selection information and the result information of the character recognition result described above via the LAN interface unit 7.
It is possible to output to another device connected on N.
In this case, the information temporarily stored in the memory 804 of the image memory unit 8 is output to the core unit 10 and the core unit 1
0 to the LAN interface unit 7 and LA
Forwarded to another device connected on N. Here, the output from the image memory unit 8 to the core unit 10, the output from the core unit 10 to the LAN interface unit 7, and the operation of transmitting from the LAN interface unit 7 to the LAN are as described in the description of each unit described above. Is.

In the image memory unit 8, (1) input of image data to the image memory unit 8 described above,
A series of instruction information regarding (2) execution of block selection processing and character recognition processing, and (3) output of processing result data from the image memory unit 8 is sent from another device connected to the LAN via the LAN interface unit 7. It is possible to accept and execute each processing accordingly. Control of execution of various processes based on the instruction information input from the outside is performed by the CPU 806. In this case, the process execution information is input to the LAN interface unit 7 from another device connected to the LAN, is input from the LAN interface unit 7 to the core unit 10, and the core unit 1
Input from 0 to the memory of the image memory unit 8. Here, the operation of inputting from the device on the LAN to the core unit 10 via the LAN interface unit 7 and inputting from the core unit 10 to the image memory unit 8 is as described in the description of each unit described above.

In the image processing system, the image information data input from the external storage device 6 connected to the scanner unit 1, the fax unit 4, and the file unit 5 which are the image information data input means included in the present system are processed. The block selection processing and the character recognition processing are executed, and the block selection processing result information and the character recognition result information obtained as a result are LA-connected via the LAN interface unit 7.
A series of processing instructions, such as outputting to another device connected on the LAN, is transmitted via the LAN interface unit 7 to the LAN.
It can be done by other devices connected above.

FIG. 33 is a flowchart showing a processing procedure according to an instruction from another device. First step S3
In 301, the instruction to execute the image information data input process is L
This is performed from another device connected to the AN to this device. The instruction regarding the image information data input process is performed by transmitting the instruction information data from the other device connected to the LAN to the image memory unit 8. The instruction information data transmitted to the image memory unit 8 is temporarily stored in the memory 804 of the image memory unit 8 and then interpreted by the CPU 806 to control the process execution in step S3302. The transfer of the instruction information data to the image memory unit 8 has been described in detail in the description of the LAN interface unit 7 and the core unit 10, and will not be repeated here.

At the following step S3302, the step S
According to the instruction of 3301, image information data input means included in the system inputs image information data to the image memory unit 8. As the image information data input means included in this system, any one of the scanner unit 1, the fax unit 4, the file unit 5 and the external storage device 6 connected thereto can be used. The input of the image information data from the scanner unit 1, the fax unit 4, the file unit 5, and the external storage device 6 connected to the scanner unit 1, the fax unit 4, the file unit 5, and the like to the core unit 10 has been described in detail in the description of the core unit 10 and thus will be omitted here.

The image information data input to the core section 10 is transferred to the image memory section 8 via the connector 1008 and then temporarily stored in the memory 804. The input of the image information data to the image memory unit 8 has been described in detail in the description of the image memory unit 8 and will not be repeated here.

Next, in step S3303, an instruction to execute the block selection process or the character recognition process for the input image data is given to this device from another device connected to the LAN. The instruction to execute the block selection process or the character recognition process is performed by transmitting the instruction information data to the image memory unit 8 from another device connected to the LAN. The instruction information data transmitted to the image memory unit 8 is temporarily stored in the memory 804 of the image memory unit 8 and then interpreted by the CPU 806 to control the process execution in step S3304. Here, the processing execution instruction can be given only for the block selection processing or only for the character recognition processing. It is also possible to instruct both of them and perform the character recognition process only on the image information data classified as the character part or the title part by the block selection process.

Further, by designating the coordinates relating to the partial area of the image information data in the instruction information data, it is possible to instruct to execute the block selection processing or the character recognition processing only for the partial area of the image information data. Is. The transfer of the instruction information data to the image memory unit 8 has been described in detail in the description of the LAN interface unit 7 and the core unit 10, and will not be repeated here.

Next, in step S3304, block selection processing or character recognition processing for the image data input according to the instruction of step S3303 is executed. Since the details of the processing have been described above, they are omitted here. Further, the result information data of the block selection process or the character recognition process obtained as a result of the process is temporarily stored in the memory 804.

Next, in step S3305, the memory 8
An instruction for outputting the result information data of the block selection process or the character recognition process temporarily stored in 04 to another device connected to the LAN is given to this device from the other device connected to the LAN. The instruction for outputting the block selection processing / character recognition processing result data is performed by the image memory unit 8 from another device whose instruction information data is LAN-connected.
This is done by sending to F. Image memory section 8
The instruction information data transmitted to the image memory unit 8 is temporarily stored in the memory 804 of the image memory unit 8 and then interpreted by the CPU 806 to control the output process execution in step S3306.

Finally, in step S3306, the result information data of the block selection processing and the character recognition processing temporarily stored in the memory 804 according to the instruction of step S3305 is output to another device connected to the LAN.
The output of data to another device connected to the LAN has been described in detail in the description of the core unit 10, and will be omitted here.

Further, in the above-mentioned image processing system, the block selection process and the character recognition process are further executed on the image information data held in another device connected to the LAN via the LAN interface section 7. , A sequence of processing instructions such as outputting the resulting block selection processing result information and character recognition result information to another device connected to the LAN via the LAN interface unit 7 via the LAN interface unit 7. Can be performed from another device connected to.

FIG. 34 is a flow chart showing the processing in this case. First, in step S3401
An instruction to execute the image information data input processing held by another device connected to the LAN via the interface unit 7 is issued to this device from the same other device connected to the LAN. Instructions regarding image information data input processing
This is performed by transmitting the instruction information data from the other device connected to the LAN to the image memory unit 8. The instruction information data transmitted to the image memory unit 8 is temporarily stored in the memory 804 of the image memory unit 8 and then the CPU 8
It is interpreted according to 06, and processing execution control in steps S3402 and S3403 is performed. still,
Since the transfer of the instruction information data to the image memory unit 8 has been described in detail in the description of the LAN interface unit 7 and the core unit 10, the description thereof is omitted here.

At the following step S3402, step S
According to the instruction of 3401, the image information data held by the other device connected to the LAN via the LAN interface unit 7 is transferred to the present system. LA
Regarding transfer of image information data from another device connected on the LAN through the N interface unit 7, the LAN is used.
Since the details of the interface section 7 have been described, the description thereof is omitted here.

Next, in step S3403, the image information data transferred to the system according to the instruction of step S3401 is input to the image memory unit 8. The details of inputting the image information data to the image memory unit 8 have been described in detail in the description of the image memory unit 8 and the core unit 10, and will not be repeated here. The image information data input to the image memory unit 8 is stored in the image memory unit 8
Is temporarily stored in the memory 804.

Since the processing of the subsequent step S3403 is the same as the processing of steps S3303 to S3306 shown in FIG. 33, it is omitted here.

[Other Embodiments] In this embodiment, a communication line to which another device is connected is a LAN (local
Although it has been described that it is an area network), this communication line may be any generally known communication line such as WAN (wide area network). It may be another device connected to the telephone line via No. 4. Further, any other device connected to the communication line may be any device as long as the device is connectable to a generally known communication line.

In this embodiment, as other devices connected via a LAN, (1) another device for transferring image information data to be processed, (2) block selection process and character recognition process are performed. Regarding the other device which is the output destination to output the executed result data, and (3) the other device which instructs such a series of operations, the example in the case of being the same device has been described. (3)
The above devices are not necessarily the same, and may be a plurality of devices connected on the LAN.

The present invention also provides a "host computer,
The invention may be applied to a system including a plurality of devices such as an interface and a printer, or to an apparatus including one device such as a copying machine. Further, it goes without saying that the present invention can be applied to the case where it is implemented by supplying a program to a system or an apparatus. In this case, the storage medium storing the program according to the present invention constitutes the present invention. Then, by reading the program from the storage medium to the system or device, the system or device operates in a predetermined manner.

[0192]

【The invention's effect】

As described above, according to the present invention, it is possible to instruct execution of predetermined image processing from another device connected to the network via the interface.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of an image forming apparatus according to this embodiment.

FIG. 2 is a cross-sectional view showing structures of a reader unit 1 and a printer unit 2.

FIG. 3 is a block diagram showing a configuration of an image processing unit in the reader unit 2.

FIG. 4 is a block diagram showing a detailed configuration of a core unit 10.

5 is a block diagram showing a detailed configuration of a fax unit 4. FIG.

FIG. 6 is a block diagram showing a detailed configuration of a file section 5.

FIG. 7 is a block diagram showing a configuration of a LAN interface unit 7.

FIG. 8 is a block diagram showing a configuration of an image memory unit 8.

FIG. 9 is a schematic flowchart showing an algorithm of block selection processing.

FIG. 10 is a flowchart showing detection of connected pixels and classification processing thereof.

FIG. 11 is a flowchart showing detection of connected pixels and classification processing thereof.

FIG. 12 is a flowchart showing detection of connected pixels and classification processing thereof.

FIG. 13 is a diagram for explaining extraction of connected components by contour line tracking.

FIG. 14 is a diagram for explaining a cutout of a rectangle that covers connected components.

FIG. 15 is a diagram showing a rectangular tree structure corresponding to connected components.

FIG. 16 is a diagram for explaining a process for checking the line width of a connected component.

FIG. 17 is a diagram showing a vertical histogram of a non-text unit.

FIG. 18 is a diagram for explaining a process of searching for a white contour in a connected component.

FIG. 19 is a diagram for explaining a process of combining white contours.

FIG. 20 is a diagram for explaining a process for combining white contours.

FIG. 21 is a diagram for explaining a process of combining white contours.

FIG. 22 is a flowchart showing a process of grouping connected pixels into rows.

FIG. 23 is a flowchart showing a process of grouping text lines into text blocks.

FIG. 24 is a diagram showing a page 90 of pixel image data.

FIG. 25 is a diagram in which connected pixels are detected for the image shown in FIG. 24.

FIG. 26 is a schematic diagram showing a processing result of grouping.

FIG. 27 is a diagram showing a block structure obtained by the processing of FIG. 23.

FIG. 28 is a diagram showing a tree structure of the obtained text.

FIG. 29 is a diagram showing a tree structure of an obtained text line.

FIG. 30 is a diagram showing a tree structure of an obtained text block.

FIG. 31 is a flowchart showing an algorithm of character recognition processing.

FIG. 32 is a diagram illustrating an example of character cutout processing by separating connected pixel components.

[FIG. 33] A LAN for issuing an instruction for performing block selection processing and character recognition processing for image information data input from image information data input means included in the present system.
It is a processing flow diagram at the time of performing from the other apparatus connected above.

FIG. 34 is a process in the case where the image information data held in another device connected to the LAN is input and an instruction for performing the block selection process and the character recognition process is given from the other device connected to the LAN. It is a flowchart.

FIG. 35 is a diagram showing a combination of non-text blocks to be combined.

[Explanation of symbols]

 1 Reader Part 2 Printer Part 3 External Device 4 Fax Part 5 File Part 6 External Storage Device 7 LAN Interface Part 8 Image Memory Part 10 Core Part 11 Hard Disk 12 LAN 13 Telephone Line

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Eiji Ohara 3-30-2 Shimomaruko, Ota-ku, Tokyo Canon Inc. (72) Inventor Yuka Nagai 3-30-2 Shimomaruko, Ota-ku, Tokyo Canon Within the corporation

Claims (6)

[Claims] 1. An interface unit for connecting to a network, a receiving unit for receiving instruction information for instructing execution of predetermined image processing from another device connected to the network by the interface unit, and the receiving unit. The input means for inputting the image information and the image processing means for processing the image information are controlled on the basis of the instruction information received by the control means so as to execute predetermined image processing on the input image information. An image processing apparatus comprising: a control unit; and a transmission unit that transmits an image on which a predetermined image process is performed by the control unit. 2. The image processing apparatus according to claim 1, wherein the predetermined image processing is block selection processing and character recognition processing. 3. The image processing apparatus according to claim 1, wherein the input unit inputs image information from another device via the interface unit. 4. A receiving step of receiving, via an interface, instruction information for instructing execution of predetermined image processing from another device connected to the network, and based on the instruction information received by the receiving step, An input step of inputting image information, an image processing step of processing the image information, and a control step of controlling the input image information to perform predetermined image processing; And a transmitting step of transmitting the image on which the image processing has been performed. 5. The image processing method according to claim 4, wherein the predetermined image processing is block selection processing and character recognition processing. 6. The image processing method according to claim 4, wherein the inputting step inputs image information from another device via the interface.