JPH0316362A - Picture communication equipment - Google Patents

Abstract

PURPOSE:To apply recognition processing of an input picture data at a high speed by deciding the recognition processing used by a recognition processing means in response to the detected coding system. CONSTITUTION:A compression system detection circuit 13 consists of the software such as a main body control section 11 to discriminate kinds of the compression coding system used for the transmission of a picture data decided by the communication protocol such as MH, MR or MMR system. Thus, the optimum recognition processing system to correct an error of an input picture data caused is selected corresponding to plural coding systems to apply the recognition processing of the input picture. When the decoded picture data is subject to character recognition, even if any transmission error exists, a high speed recognition processing is attained.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to an image communication device, and particularly to image data encoded by one of a plurality of encoding methods via a line or a predetermined input means. The present invention relates to an image communication device that inputs an image, decodes the input encoded image data, and recognizes the decoded image. [Prior Art] Facsimile devices that transmit image data via telephone lines, digital lines, etc. have been known for some time, and techniques for character recognition of transmitted image data on the receiving side are also known. Conventional devices that recognize transmitted image data usually decode the transmitted image data, develop it into binary image data, and then perform character recognition. [Problem to be solved by the invention] However, in data compression methods such as MR (Modified Read) used in G3 and G4 facsimiles, if a transmission error occurs when image data is transmitted over a line, subsequent Since the lines are disrupted up to three lines at most, there is a problem in that characters are often unrecognizable or misrecognized during character recognition.

It may be possible to perform character recognition by interpolating image data in anticipation of transmission errors of a few lines, but this would complicate the character recognition algorithm and take time to process the recognition process. A problem occurs. An object of the present invention is to solve the above problems and to enable high-speed recognition processing even if there is a transmission error when character recognition is performed on decoded image data.

[Means for Solving the Problems] In order to solve the above problems, in the present invention, images encoded by one of a plurality of encoding methods via a line or via a predetermined input means are provided. A process for correcting image data errors that occur in response to multiple encoding methods for input image data in an image communication device that inputs data, decodes the input encoded image data, and recognizes the decoded image. A recognition processing means that can execute a plurality of recognition processes including We adopted a configuration that determines recognition processing.

[Operation] According to the above configuration, it is possible to perform recognition processing on an input image by selecting the recognition processing method most suitable for correcting errors in input image data that occur in response to a plurality of encoding methods. can. [Example] The present invention will be described in detail below based on the example shown in the drawings. Figure 1 shows the configuration of an image communication device that employs the present invention.
In FIG. 1, reference numeral 19 is a communication line such as a public telephone network, which is connected to the communication control section 14. The communication control unit l4 is composed of a known network control device, a modem, and an encoding/decoding unit, and G3. Image data is sent and received using the G4 facsimile procedure. The encoding/decoding unit of this communication control unit l4 is a known MH. It compresses and expands image data using encoding methods such as MR and MMR.

Image data transmitted and received via the communication control unit l4 is input and output to and from the line using the image memory l2 as a buffer. In this embodiment, the communication control unit 14 and the image memory l
A detection circuit 13 for identifying the compression method of the image data to be sent and received, in particular, to be received, is connected between the two. The compression method detection circuit 13 is configured by software such as the main body control unit 11, which will be described later, and detects the type of compression encoding method used for transmitting the image data determined in the communication procedure, such as MH, MR, or MMR method. identify

The image memory l2 includes a printer 15 consisting of a laser beam printer for outputting images, and a printer 15 for inputting images.
A scanner 16 consisting of a CD line sensor and a document conveyance system, etc., and a hard disk 17 used for storing a large amount of image data or as a virtual storage device are connected. Reference numeral 1l denotes a control unit composed of a microprocessor, etc., which mainly controls the entire communication process. The communication control program of the main body control unit 1l is stored in the ROMI a.

In this embodiment, a character recognition unit l8 is further provided,
This character recognition section 18 is composed of a known microprocessor and its peripheral elements. A ROM 18a storing recognition control procedures is stored in the character recognition unit l8.

The character recognition unit l8 recognizes the decoded received data stored in the image memory 12 using a known character recognition procedure, converts it into a character code corresponding to the characters included in the image data, and stores it on the hard disk. 17, or output to other processing devices such as a computer system via a human output interface (not shown). In this embodiment, the character recognition algorithm of the character recognition unit l8 is not one type, but a plurality of algorithms are prepared depending on the compression method of the received data (detected by the encoding method detection circuit l3). In the control described later, MMR. It is assumed that the MR method is used, and in particular, when the MMR method cannot be used due to the possibility of errors occurring on the line, the error-prone MR method is used, and the K parameter during compression processing is Value: (for example, K
= 2 or K = 4), a recognition procedure including different error correction processing is prepared. Furthermore, if the MMR system can be used, a retransmission procedure can be used even if there is a transmission error on the line, so an algorithm that does not take error correction into account is used during character recognition. These three different algorithms are ROM
Store in l8a.

Next, differences in transmission errors caused by differences in compression methods used for transmission will be explained.

As mentioned above, the MMR method does not require consideration of transmission errors, but the MR method used in G3 facsimile machines, etc.
The encoding method is a two-dimensional encoding method, and when a data error occurs in a certain encoding line, the error affects the following lines.

For example, the lines 21 and 23 in FIG. 2 are l-dimensional encoded lines, and the lines 22 and 24 following these lines are two-dimensional encoded lines that are encoded while referring to the lines 2l and 23. Figure 2 shows a case where the K parameter of MR encoding is 2. For example, if an error occurs on line 2l, all the data on line 22 that refers to this will generate an error, and in the end, 2 lines will have a transmission error. It is influenced by. K parameter is usually 2 to 4, but K=
In the case of 4, 4 lines are similarly affected by the error.

FIG. 3 shows how images are actually affected by transmission errors. Figure 3 shows the same 1 to 4 rows.
An example is shown in which an image including the number 5 and two vertical bars sandwiching it is transmitted.

In the first line, no error occurs and the image is received normally, but in the second line, an error occurs in the line in the center of the characters, and one line is missing. In the third line, an error occurs in the one-dimensional encoded line, and the following two lines that refer to that line are also affected, and the center part of the character is distorted.

In the fourth line, an error occurred as in the third line, so this error line was ignored and the images before and after it were combined. By deleting this error line, numbers such as "1" are not affected, but some of the characters in numbers "2" and "4" are cut off.

If an image is transmitted using a two-dimensional compression method, such as an MR encoding method, and an error occurs on the line, the image output by the receiving device will be the one shown in lines 2 to 4 in Figure 3. You will have to sweat one of these two. When recognizing a received image such as the lines 2 to 3 in FIG. 3 using the character recognition unit l8, it is especially important to avoid missing lines such as the lines 2 or 3 in FIG. If we know the maximum number of lines that can continue to fail, we can use an appropriate character recognition algorithm.

Figure 4 shows an overall flowchart of the communication and character recognition procedures for the device shown in Figure 1. In step S1 in FIG. 4, the communication control unit l4 executes a known communication procedure and receives image data from the line l9. The received image data is sequentially decoded within the communication control unit,
The expanded binary image data is stored in the image memory 12. Since the compression method of the image data is usually known through the communication procedure, this compression method is detected in step S2 by the encoding method detection circuit l3. Step S3 is
It shows the storage of decompressed image data into the image memory l2.

In step S4, it is determined whether all the image data has been received, and when the reception of the image data is completed, step S4 is performed.
At step 5, the main body control section 11 notifies the character recognition section 18 that the reception of image data has ended. At the same time, the compression method of the received image data detected by the encoding method detection circuit 13 is notified. In step S6, the character recognition unit l8 selects an optimal algorithm suitable for the compression method used when transmitting the decompressed image data stored in the image memory l2, and performs recognition processing in step S7.The recognition result is as follows. Step S8
The main body control unit ll is notified. Figure 5 shows the recognition processing procedure of the character recognition unit l8. Here, as mentioned above, it is assumed that the MMR method and two types of MR methods with K parameters of 2 and 4 are used during image transmission, and the encoding method detection circuit l3 discriminates between these three compression methods, The main body control section 11 notifies the character recognition section 18 of the detected compression method.

Step S11 in FIG. In the determination step of Sl3, 515, the compression method of the image data notified from the main body control unit 11 is identified. Step SL! Then the compression method is M
Is it the MR method?In step S13, the K parameter is 2.
R method or MR with K parameter of 4 in step 515
The method is determined. If any of the determination steps is affirmed, step 51
2, S14. Proceed to S16. In the case of the MMR method, there is no need to take transmission errors into account as described above, and therefore, in step SL2, a character recognition algorithm that does not include error correction processing and is stored in the ROM 18a is selected.

In the case of the MR method where the K parameter is 2, the number of consecutive lines affected by the error is 2, so step 51
In step 4, select a character recognition algorithm that includes relatively simple error correction processing. In the case of the MR method where the K parameter is 4, the number of consecutive lines affected by errors is 4, so in step S16 a character recognition algorithm that includes more powerful error correction processing than that in step S14 is selected.8 On the other hand, in step ' S11. If both Sl3 and S15 are negative, the compression method is determined to be unrecognizable because it cannot be identified, and the main body control unit 11 is notified to that effect. When steps S12, S14, and Sl6 are completed, recognition processing of the binarized image data stored in the image memory 12 is executed using the character recognition algorithm selected in step 51B.

In addition, step SL? Now, the image data stored in the image memory l2 is converted into an error. Characters may be recognized without any correction.

According to the above embodiment, when character recognition is performed on decoded image data, the encoding compression method used to transmit the image data is identified, and an error suitable for the error characteristics corresponding to the encoding method is detected. Character recognition is performed by selecting an algorithm that includes correction processing. Therefore, error correction suitable for transmission errors that may occur can be performed to ensure reliable and as fast character recognition processing as possible. Although character recognition has been exemplified above, "character" has a broad meaning, and of course includes various marks, symbols, symbols, figures, etc. It goes without saying that the above configuration is not limited to facsimile image communication, but can be applied to two types of image data transmission. Furthermore, the image communication method is not limited to the facsimile method, and it goes without saying that similar control is possible when processing encoded image data input from an image reading unit or the like. [Effects of the Invention j As is clear from the above, according to the present invention, image data encoded by any one of a plurality of encoding methods is input via a line or via a predetermined input means, In an image communication device that decodes input encoded image data and recognizes the decoded image, multiple recognition processes including processing to correct image data errors that occur in response to multiple encoding methods of input image data are performed. A recognition processing means capable of executing the process and a means for detecting the encoding method of input image data are provided, and a recognition process to be used by the recognition processing means is determined according to the encoding method detected by the detection means. This configuration allows the recognition processing of the input image to be performed by selecting the optimal recognition processing method to correct errors in the input image data that occur in response to multiple encoding methods. Even if there is an error, it has the excellent effect of performing correction processing appropriate to the characteristics of the error and allowing high-speed recognition processing of input image data.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of a facsimile device adopting the present invention, FIG. 2 is an explanatory diagram showing the encoding and compression method, and FIG. 3 is an explanatory diagram showing an image in which a transmission error has occurred.
FIG. 4 is a flowchart showing the image communication and character recognition procedure, and FIG. 5 is a flowchart showing the character recognition processing procedure of the character recognition section of FIG. 11...Main body control unit l2...Image memory l3.
...Encoding method detection circuit 14...Communication control unit l5...Printer 16...Scanner 17...
・Hard disk l8...Character recognition section l9...Line 3rd figure Qian V Zhanshu'Il+White flow cage I-Aki 4th figure 5th figure

Claims (1)

[Claims] 1) Inputting image data encoded by one of a plurality of encoding methods via a line or a predetermined input means, and decoding the input encoded image data. In an image communication device that recognizes encoded images, a recognition processing means capable of executing a plurality of recognition processes including a process of correcting image data errors that occur in response to a plurality of encoding methods of input image data; An image communication device comprising means for detecting an encoding method of input image data, and determining a recognition process to be used by the recognition processing means in accordance with the encoding method detected by the detecting means.