JPH04342089A - Character input procedding method - Google Patents

Abstract

PURPOSE:To efficiently segment the handwritten sentence from a handwritten document by funding a line head position coordinate and a line range to determine a line position, and subsequently, finding a character head position in each line and a character range to determine a character position, and executing the character conversion and the character recognition. CONSTITUTION:A processing part 3 determines, first of all, a line position by finding a line head position coordinate and a line range from input data of a handwritten sentence (character) read by an OCR 1. Subsequently, the processing part 3 determines a character position by finding a character head position coordinate in each line and a line range. That is, the processing part 3 finds a start position and a range of the head character of a certain line, first of all, a first line, executes it with regard to all characters of this line and finds its one character range. The processing part 3 executes the character conversion (expanding or compression processing) to these dot image patterns in each character range, executes the character recognition with regard to image data obtained in such a way, and also, repeats the processing until that time with regard to each line. In such a way, the sentence can be segmented efficiently.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character input processing method, and more particularly to a character input processing method that makes it possible to efficiently extract handwritten text from a handwritten document.

0002

2. Description of the Related Art Conventionally, as a character extraction method in this type of character input processing device, for example,
Methods disclosed in Japanese Patent Application Laid-open No. 8362 and Japanese Patent Application Laid-Open No. 56-44972 are known. All of these techniques indicate a technique for searching within a line area of a predetermined width for typed or printed characters and cutting out each typed or printed character.

0003

Problems to be Solved by the Invention The above-mentioned prior art does not take into consideration the extraction of handwritten sentences from handwritten documents. This is because when extracting handwritten text from a handwritten document, everything such as character size and line spacing varies widely, and it is extremely difficult to do this automatically. The present invention has been made in view of the above circumstances, and its purpose is to solve the above-mentioned problems in the conventional technology and to provide a character input that enables efficient extraction of handwritten sentences from handwritten documents. The purpose is to provide a processing method.

0004

[Means for Solving the Problems] The above object of the present invention is to
In a character input processing device that has a CR (optical character reader), a keyboard, and a display and inputs handwritten text (characters), first, from the input data of the handwritten text (characters), line start position coordinates and line range are determined. The line position is determined by determining the line position, then the character position is determined by determining the character start position and character range within each line, character conversion is performed on the obtained image data, and then character recognition is performed. This is achieved by a character input processing method.

[0005]

[Operation] In the character input processing method according to the present invention, from the input data of handwritten text (characters), the line position is determined by first finding the line start position coordinates and line range, and then the character start position within each line. The system determines the character position by determining the character range, performs character conversion on the obtained image data, and then performs character recognition, making it possible to automatically input characters from any handwritten character. becomes possible.

[0006]

Embodiments Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 2 is a block diagram of a character input processing device showing an embodiment of the present invention. In the figure, 1 is an OCR (optical character reading device) for reading handwritten text (characters);
2 is a keyboard for changing and correcting characters read by the OCR 1; 3 is a processing unit that performs processing such as character recognition and type conversion, which will be described later; 4 is an input buffer for buffering image data from the OCR 1; Reference numeral 5 indicates a display section that displays text after processing such as character recognition and type conversion, and 6 indicates the same printer. FIG. 1 shows the processing unit 3
FIG. 3 is a conceptual diagram of the process of determining the line range and single character range of handwritten text (characters) read by CR1. The handwritten text (characters) read by the CR1 is input to the input buffer 4 as a dot image pattern of "0" and "1", although FIG. 3 shows an example of a horizontally written document (step 11). Next, regarding the data of the dot image pattern input to the input buffer 4, the processing unit 3 performs the following:
The range for one line and the range for one character are determined (steps 12 and 13).

The determination of the range for one line and the range for one character described above is performed as shown below. That is, first, Figure 3
As shown in , at each position, while shifting the Y coordinate by +1 from the reference point (X0, Y0) in the X direction and Y direction,
Find the Y coordinate Ya at which the logical sum of the data from the X coordinates X0 to Xn is "1". Note that the above Xn is a position predetermined depending on the size of the paper. Ya becomes the starting point of the first line character range. Next, find the final Y coordinate of the first line character range. This means that while moving the Y coordinate by +1 from the first Y coordinate (Ya) of the first line character range found earlier, the logical sum of the data from X coordinates X0 to Xn becomes a continuous "0" Y Yb is determined by finding the start of the coordinates. Through the above processing, the line range for the first line of the text is determined. Next, the start position and range of the first character in the first line determined by the process described above are determined. The start position of the first character is the start of the X coordinate where the logical sum of the data from the Y coordinate Ya to Yb found earlier as the line range continuously becomes "1"; here, Xb is found. Also, in order to find the final X coordinate of the character range, while moving the X coordinate by +1 from Xb found earlier, find the Just find the beginning Xc. Through the above processing, the position and range of the first character in the first line are determined. Thereafter, the same operation is repeated to determine a one-character range of all the characters in the first line.

During the above operation, if the interval from the final coordinate of the previous character to the first coordinate of the next character in the same line exceeds a specified value, this is regarded as a space, and the first coordinate of the next character is shifted forward. , delete the space. Since the size of each handwritten character is different, the dot image pattern within one character range is divided into a predetermined frame size (for example, 2
4x24 dot image pattern) is expanded or compressed, and then character recognition is performed (step 14,
15). For the second and subsequent lines, in the same way as the first line, the line range determination, character start position and range determination, image data expansion/compression processing, etc. (these processes are collectively referred to as "character conversion") are performed. After this, character recognition is performed on the obtained image data, and this is repeated until the final character line (steps 16 and 17). Regarding line spacing,
Similar to deleting empty space within a line, if the starting coordinate of the next line from the final coordinate of the previous line is greater than or equal to a specified value, the starting coordinate of the line is moved and the blank line is deleted.

Next, the text is formatted. If there is an extra space after the last character on a line, and the last character is not a period, move the next line's characters by the specified number of digits to the previous line. Note that if the last character is a period, the character on the next line is moved by one digit. By repeating this operation, the text is formatted. In addition, in the above-mentioned character conversion process, for characters that cannot be processed because the input characters are unclear or complicated, corrections can be made by displaying them as image data as they are or by displaying them in a blinking manner when displaying them on the display section 5. It's good to make it easier. According to the embodiment described above, it is possible to automatically convert a handwritten manuscript into a fixed form sentence format. It should be noted that the above-mentioned embodiment shows an example of the present invention, and it goes without saying that the present invention should not be limited thereto.

0010

[Effects of the Invention] As described above in detail, the present invention has the remarkable effect of realizing a character input processing method that makes it possible to efficiently extract handwritten sentences from handwritten documents. be.

[0011]

[Brief explanation of the drawing]

FIG. 1 is a flowchart of processing by a processing unit 3 of a character input processing device that is an embodiment of the present invention.

FIG. 2 is a configuration diagram of a character input processing device according to an embodiment.

FIG. 3 is a conceptual diagram of processing in the embodiment.

[Explanation of symbols]

1: OCR, 2: Keyboard, 3: Processing unit, 4: Input buffer, 5: Display unit, 6: Printer, 11 to 17: Processing steps.

Claims (1)

[Claims] 1. A character input processing device for inputting handwritten text (characters), which includes an OCR (optical character reader), a keyboard, and a display, wherein the handwritten text (characters) is
From the input data of (characters), first determine the line position by finding the line start position coordinates and line range, then determine the character position by finding the character start position and character range within each line, and the resulting image data A character input processing method characterized in that character recognition is performed after character conversion is performed.