JPH04336683A - Method and device for correcting character recognition data written vertically - Google Patents

Abstract

PURPOSE:To reduce user's burden for correction by utilizing the character information of a character recognition device and displaying an original picture after converting its vertical writing into horizontal writing. CONSTITUTION:An original 10 written vertically is read by a scanner 11 and obtains picture data. This picture data is stored in a picture data storage part 1. A character recognition part 12 segments the picture data in rectangular for each character and separates into character frames, character information composed of the coordinate, size, character code of the character frame is generated to be stored in a character information storage part 2. A one character picture data reading means 5 capable of taking out one character of picture data from the picture data storage part 1 with the coordinate and size of the character frame designated and a vertically written character decision means 6 deciding that a character is a vertically written one which is inconvenient for horizontal writing, are provided. By using character information, the vertically written original picture is converted into horizontal written one. Thus, it can be easily compared with the character code of the recognition result at the time of the correction.

Description

[Detailed description of the invention]

0001

[Industrial Application Field] The present invention is a character recognition device that converts a document image into horizontal writing and displays it so that when a user modifies a vertically written original, it is easier to compare the character code of the recognition result. The present invention relates to a method and device for correcting written character recognition data.

0002

2. Description of the Related Art Conventionally, character recognition data is corrected by displaying a vertically written original image as it is written vertically.

0003

[Problems to be Solved by the Invention] However, since character data is normally written horizontally on a computer, it is difficult to see when comparing it with an image written vertically, causing inconvenience to the user.

[0004] An object of the present invention is to provide a character recognition device that:
An object of the present invention is to provide a method and device for correcting vertically written character recognition data, which reduces the user's correction burden by converting a vertically written original image into horizontally written and displaying the image.

[0005]

[Means for Solving the Problems] The present invention cuts each character into rectangles from vertically written original image data read by a scanner, separates them into character frames, recognizes characters for each character frame, and coordinates the character frame. A character recognition device that generates character information consisting of , size, and character code is characterized in that the character information is used to convert a vertically written original image into horizontally written image and display the converted image; Cuts each character into a rectangle from the vertically written original image data read by a scanner, separates it into character frames, recognizes the characters for each character frame, and generates character information consisting of the coordinates of the character frame, size, and character code. The character recognition device includes an image data storage section that stores document image data, a line management means that manages line digits to be processed, a character information storage section that stores the character information, and a line information storage section that stores the character information. Character information reading means for reading out character information for one character pointed to by the digit management means, single character image data reading means for retrieving one character's worth of image data from the image data storage section by specifying the coordinates and size of a character frame, and when writing horizontally, a vertical writing only character determination means for determining whether it is an inconvenient vertical writing only character from a character code; a horizontal writing image conversion means for converting an image of a vertical writing only character from vertical to horizontal;
A display position management means for managing the display position so that it is displayed horizontally, and 1 for displaying a single character image at an arbitrary display position.
It is characterized by having a character image display means.

[0006]

[Operation] In the present invention, first, character information for one character is retrieved from the character information storage section by the character information reading means by specifying the line position managed by the line position management means.

Next, by specifying the coordinates and size of the character frame of the retrieved character information, the single character image data reading means retrieves image data for one character from the image data storage section. Next, the vertical writing only character determination means determines from the character code whether the character is a vertical writing only character that would be inconvenient when written horizontally. If it is a vertical writing only character, the horizontal writing image conversion means will
Converts the image of one character horizontally and vertically.

1 at the display position specified by the display position management means.
A character image display means displays a single character image. When the display is finished, the display position management means changes the display position to the right-hand position so that the text is written horizontally.

As described above, one character is converted to horizontal writing and displayed.

The line column management means changes the line column so that it points to the character in the next column. However, if it is at the end of a line, it points to the first column of the next line. If there is the next character, repeat the above process.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is an example of a block diagram of a vertical writing character recognition data correction apparatus according to the present invention. Vertical manuscript 1
0 is read by the scanner 11 to obtain image data. The image data is stored in the image data storage section 1. Character recognition section 1
In step 2, image data is cut into rectangles for each character, separated into character frames, characters are recognized for each character frame, character information consisting of the coordinates of the character frame, size, and character code is generated, and is stored in the character information storage unit 2. Pay. An example of the structure of character information is listed in character information sample 14. With this structure, character information corresponding to the number of recognized characters is stored in the character information storage section 2. The line digit management means 3 holds the line digits to be processed,
After processing one character, it points to the next character. The next character is the next column, but at the end of the line it is the first column of the next line. The character information of one character of the line point indicated by the line point management means 3 is read out from the character information storage section 2 by the character information reading means 4. A single character image data reading means 5 reads image data for one character from the image data storage section 1 based on the coordinates and size of the character frame in the character information. The vertical writing only character determining means 6 judges from the character code in the character information whether or not the character is a vertical writing only character (for example, a parenthesis or a long sound) that is inconvenient even if it is arranged horizontally as it is. If the character is a character exclusively for vertical writing, the horizontal writing image converting means 7 converts it into an image suitable for horizontal writing. In most cases, the solution is to rotate the camera 90 degrees to the left (counterclockwise). However, long sounds and hyphens are
In order to avoid difficulty in distinguishing characters, including the kanji 1, it is necessary to take measures such as changing hyphens to straight lines with a width of 1 dot. The display position management means 8 first adjusts the display position so that if the characters are on the same line, they are lined up to the right of the previous character. Since the size of the character frame is not square, trying to display it in the same position will not work, and delicate adjustments will be required. For example, you can adjust the balance between characters so that the spacing between characters is not too wide or too narrow, or calculate the average value of the coordinates and size of the character frame and then adjust it so that the top and bottom of the character string are as aligned as possible. ing. Furthermore, if the text appears to be overflowing to the right edge of the screen, it is possible to make adjustments by further reducing the spacing between characters or by wrapping the display. When moving to the next line, change the display position so that it looks like a new line. The single character image display means 9 displays a single character image on the CRT 13 at the display position designated by the display position management means 8. Note that in the single character image display means 9, the unit of display position is dot unit, so for example, if the video RAM of the computer is in byte unit, bit shifting (justifying) is performed. Further, an example of a screen display that is an effect of the present invention is shown in FIG.

FIG. 2 is an example flowchart of the present invention. In F1, original image data is input from the scanner. At F2, character recognition is performed to generate character information and store it in the character information storage section. In F3, initial setting is made to row=1 and column=1. In this embodiment, an image of one character is displayed in one process, and the process continues in a loop until all characters are exhausted. This line column indicates the character currently being processed. Note that the line column is updated in F12 to F15, and the end of the line is checked. At F4, the character information of the character to be processed is read from the character information storage unit 2 onto the memory. F5
Now, from the coordinates and size of the character frame, image data storage section 1
Load an image of one character from into memory. At F6, the character code is checked to see if it is a character exclusively for vertical writing. In the case of vertical writing only characters, it is inconvenient to arrange them horizontally as they are, so use F7 to convert them to horizontal writing images. 1 for F8-F10
Determines the display position of the character image. F8 checks whether it is the first character of the line. If it is at the beginning of a line, press F9 to find the display position of a carriage return line feed style. Generally, the left end position of the previous line is taken as the digit position, and the line position is determined using the line spacing as per the image data. F10
In cases other than the beginning of a line, the column position is determined so that the character is aligned horizontally to the right of the previous character. However, as mentioned above, subtle adjustments are required. In F11, an image of one character is displayed at the display position thus determined. F12 to F15 manage line digits. First, increment the digit with F12. Next, use F13 to check whether it has exceeded the end of the line, and if so, use F14 to reset it to the first column of the next line. However, if the line exceeds the last line, one character image of all recognized characters has been displayed, and the process ends. If not, repeat the process from F4 onwards.

FIG. 3 shows a character frame cut out from image data 30. A rectangle surrounding a character is one character frame. The description of the X coordinate, Y coordinate, X size, and Y size of the character frame in the character information is as shown in FIG. Although the example of character information is given in 15 of FIG. 1, the units are all dots. In this embodiment, the upper left end of the image data is set as the origin (0,0). Furthermore, the origin of the line column (first line, first column) is the upper right corner, and the rightmost line is the first line. Note that character data refers to character codes (text data, JIS codes) generated by character recognition.

FIG. 4 shows an example of a screen display showing the effects of the present invention. The vertically written image data 30 shown in FIG. 3 is rearranged into horizontally written image 40 and displayed as an image. The regular characters are images as they are, and are arranged side by side while adjusting the character spacing and line spacing. To give an example of adjusting the character spacing, in (03) in FIG. 4, the character spacing will be lengthened if the character spacing on the left and right sides of the parentheses is increased, so the character spacing is slightly reduced. However, characters only for vertical writing, such as () and -, are rotated 90 degrees to the left to convert them into images for horizontal writing, and then arranged horizontally. The hyphen (-) has been intentionally changed to a thin line one dot wide so that it can be distinguished from the number 1. Note that below the horizontally written image 40, character recognition results 41 are displayed in parallel.

[0015]

[Effects of the Invention] According to the method and apparatus for correcting vertically written character recognition data of the present invention, the character recognition results of vertically written manuscripts can be checked and
On the correction screen, images that were originally written vertically can be displayed in a pseudo-horizontal manner. Normally, character data on computers is displayed horizontally, so users have a habit of having difficulty checking the vertical direction. Therefore, when the character string of the recognition result is displayed in parallel with the image for the user to check and modify, horizontal-to-horizontal matching makes a lot of sense.
It is very easy to see for users, reducing fatigue and stress, and making it easier to check and correct recognition results.

[Brief explanation of drawings]

FIG. 1 is a configuration block diagram of a vertical writing character recognition data correction device according to the present invention.

FIG. 2 is a flowchart of the device according to the invention.

FIG. 3 is a diagram showing an example of image data obtained by the apparatus according to the present invention.

FIG. 4 is a diagram showing an example of a screen display by the device according to the present invention.

[Explanation of symbols]

1: Image data storage section 2: Character information storage section 3: Line management means 4: Character information reading means 5:1 character image data reading means 6: Character determination means for vertical writing only 7: Horizontal writing image conversion means 8: Display position management means 9:1 character image display means 10: Vertical manuscript 11: Scanner 12: Character recognition section 13:CRT 14: Text information sample

Claims (2)

[Claims] Claim 1: From vertically written original image data read by a scanner, each character is cut into rectangles and separated into character frames, characters are recognized for each character frame, the coordinates of the character frame, size,
A method for correcting vertically written character recognition data, characterized in that, in a character recognition device that generates character information consisting of a character code, a vertically written original image is converted into horizontally written image and displayed using the character information. 2. Cut out each character into a rectangle from vertically written original image data read by a scanner, separate it into character frames, recognize the characters for each character frame, and calculate the coordinates of the character frame.
A character recognition device that generates character information consisting of a size and a character code, comprising: an image data storage section that stores the document image data; a line management means that manages line columns to be processed; a character information storage section for reading out character information of one character pointed to by the line management means from the character information storage section; 1-character image data reading means that can read out image data for 1 character, vertical writing-only character determination means that determines from the character code whether it is a vertical writing-only character that would be inconvenient to write horizontally, and a horizontal writing image that converts the image of a vertical writing-only character from vertical to horizontal. a conversion means;
A display position management means for managing the display position so that it is displayed horizontally, and 1 for displaying a single character image at an arbitrary display position.
A vertically written character recognition data correction device characterized by having a character image display means.