JPH0576671B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention relates to a character recognition device that recognizes printed and handwritten characters from newspapers, magazines, etc., and converts them into an amount of information such as a JIS code.

Conventional technology Conventional character recognition devices target documents with clear character spacing, that is, documents whose absolute position on the paper is known in advance, and do not impose restrictions on the target documents. Ta. To solve this problem, we cut out the character string to be recognized from the input document into a rectangle with width W and height H, and confirmed that the ratio of the length to width of the character is approximately 1. It was used to extract individual character patterns from a string of characters. (For example, Akiyama, Naito, and Masuda “Method for extracting individual characters from vertically and horizontally written documents” IEICE Technical Report
PRL83-7 Published by the Institute of Electronics and Communication Engineers) Problems to be solved by the invention However, in reality, the aspect ratio of characters is “1”.
Many of them are not close to that, and the method of cutting out individual characters based on the height of the character string causes errors in cutting out individual characters.

SUMMARY OF THE INVENTION An object of the present invention is to provide a character recognition device that can extract individual characters from a character string and perform character recognition even for characters whose aspect ratio is not close to "1".

Means for Solving the Problems The character recognition device of the present invention includes an image input section into which an image including characters to be recognized is input, and a character string to be recognized from the image inputted by the image input section with a width W, A character string cutout part that is cut out with a rectangle with a height H, and a character part whose histogram value is "1" or more by scanning the rectangle in a direction perpendicular to the character string direction to obtain a histogram of pixels that form a character. A sub-character pattern extracting section extracts a sub-character pattern composed of consecutive character parts in , and a width W _i of the sub-character pattern obtained in the sub-character pattern extracting section |W _i -H|
Find the maximum value W _j from W _i that satisfies ≦α (α: constant), set W _j as the standard width W _s of the character pattern, and combine by combining adjacent sub-character patterns among the sub-character patterns. an individual character pattern extraction unit that determines the combined sub-character pattern as an individual character pattern when the width W _p of the sub-character pattern obtained satisfies |W _p −W _s |≦β (β: constant);
The present invention is characterized in that it includes a recognition section that calculates the characteristics of the character pattern obtained by the individual character pattern extraction section and compares the characteristics with a dictionary to extract recognition candidate characters.

Function With this configuration, a histogram is obtained by scanning a rectangular character string in a direction perpendicular to the character string direction, character breaks are detected from the histogram, and sub-character patterns, which are the constituent elements of the character pattern, are determined. and set the maximum width of the sub-character pattern in the character string as the reference width of the character pattern,
Individual character patterns can be extracted by combining sub-character patterns based on the reference width, and even characters whose aspect ratio is not close to "1" can be accurately extracted and recognized.

Embodiment Hereinafter, an embodiment of the present invention will be described based on FIGS. 1 to 5.

FIG. 1 shows a block diagram of an embodiment of a character recognition device according to the present invention. Reference numeral 1 denotes an image input unit which scans an image containing characters to be recognized, inputs the image as a binary signal, and stores it in the image memory 2. 3 is a character string cutting section which scans the image memory 2 and cuts out a character string into a rectangular shape. Reference numeral 4 denotes a sub-character pattern extracting section which scans the character string cut out by the character section cutting section 3 perpendicular to the column direction to obtain a histogram of the character section and extracts sub-character patterns which are constituent elements of the character pattern. Reference numeral 5 denotes an individual character pattern extraction section which determines individual character patterns from the combination of sub-character patterns extracted by the sub-character pattern extraction section 4. Reference numeral 6 denotes a recognition unit, which calculates the stroke and other features of each character pattern extracted by the individual character pattern extraction unit 5, compares it with the character features registered in advance in the dictionary 7, and selects the most similar character as a recognition candidate. Characters. A display section 8 displays the recognition results obtained by the recognition section 6.

Regarding the character recognition device configured in this way,
The configuration will be described in detail by listing the input images shown in FIG. 2 in a row.

An image as shown in FIG. 2 input from the image input section 1 is binarized and stored in the image memory 2. The character string cutting unit 3 extracts character strings whose absolute positions are predetermined from the input image stored in the image memory 2 into a rectangle R as shown in FIG. 3a.
Cut it out. Next, the sub-character pattern extraction unit 4 scans the character string cut out in the rectangle R in the column direction and the vertical direction to obtain a histogram of the character part as shown in FIG. Then, the width W _i (i=1, 2, . . . , 8) of each sub-character pattern is determined. Sub-character pattern cut out in Figure 3c
P _s1 , P _s2 , ..., P _s8 are shown. In the individual character pattern extracting section 5, the width W _i of the sub-character pattern is determined from among the sub-character patterns extracted by the sub-character pattern extracting section 4 to be the height H of the character string cut out by the rectangle R and |W _i −H| Find the maximum value W _i among W _i that satisfies the condition of ≦α (α: constant), and set the standard character pattern as P _j
Let W _j be the standard width of the character pattern. For example, in FIG. 3b, W ₁ is the maximum, and the reference character pattern is P ₁ . Furthermore, adjacent sub-character patterns are combined, and the sub-character pattern width W _i and the width between sub-character patterns b _i satisfy the condition | 〓 ⁱ W _i + 〓 ⁱ b _i −W ₁ |≦β (β: constant) In this case, adjacent subpatterns are combined into an individual character pattern, and an individual character pattern
P ₁ , P ₂ , ..., P ₆ are determined as shown in FIG.
The recognition unit 6 determines whether M or more individual character patterns P _i obtained by the individual character pattern extraction unit 5 are continuous in the direction indicated by the arrow in FIG. 5b, including the pixel of interest, and sets a direction code. , strokes are extracted by examining the connectivity of each pixel for each direction code, and feature quantities such as the number, position, and length of strokes are extracted. FIG. 5a shows the stroke extraction result of the character ``sent''. The extracted feature amount is compared with the feature amount registered in the dictionary 7, and the most similar character is set as a recognition candidate character and displayed on the display unit 8.

Effects of the Invention As explained above, the character recognition device of the present invention includes an image input section, a character string cutting section, a sub-character pattern extraction section, an individual character pattern extraction section, and a recognition section, and a character string cut out in a rectangular shape. scan in the direction perpendicular to the character string direction to obtain a histogram, detect character breaks from the histogram to determine sub-character patterns that are constituent elements of the character pattern, and calculate the maximum width of the sub-character patterns in the character string. Since the standard width of the character pattern is used as the standard width, and individual character patterns are extracted by combining sub-character patterns based on the standard width, when extracting individual character patterns from the character string to be recognized, ”, it is possible to accurately extract individual character patterns even if they are not close to ``, and the accuracy of character recognition can be improved.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an embodiment of the character recognition device of the present invention, FIG. 2 is a diagram showing an example of an input image, FIG. 3 is an explanatory diagram of the operation of cutting out sub-character patterns from a character string, and FIG. 4 5 is an explanatory diagram of the result of cutting out individual character patterns, and FIG. 5 is an explanatory diagram of the character recognition method. DESCRIPTION OF SYMBOLS 1... Image input part, 2... Image memory part, 3... Character string extraction part, 4... Sub character pattern, 5... Individual character pattern extraction part, 6... Recognition part, 7... Dictionary, 8
...Display section.

Claims (1)

[Claims] 1: an image input unit that inputs an image containing characters to be recognized; a character string cutting unit that cuts out a character string to be recognized from the image input by the image input unit into a rectangle with a width W and a height H; Scan in the direction perpendicular to the character string direction to obtain a histogram of pixels forming a character, and extract a sub-character pattern consisting of consecutive character parts in character parts whose histogram value is "1" or more. a sub-character pattern extraction section and a width W _i of the sub-character pattern obtained in the sub-character pattern extraction section;
W _i that satisfies |W _i −H|≦α (α: constant)
Find the maximum value W _j from among them, use W _j as the reference width W _s of the character pattern, and combine adjacent sub character patterns among the sub character patterns, so that the width W _p of the combined sub character pattern is |W _p − _Ws
An individual character pattern extracting unit that uses the combined sub-character pattern as an individual character pattern when ｜≦β (β: constant) is satisfied; and an individual character pattern extraction unit that calculates the characteristics of the character pattern obtained by the individual character pattern extraction unit and A character recognition device that includes a recognition unit that extracts recognition candidate characters by comparing them with a dictionary.