JPH0584553B2 - - Google Patents

Description

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

BACKGROUND ART Conventional character recognition devices target documents with clear character spacing, that is, documents whose absolute position on a sheet of paper is known in advance, and limits the target documents.

In order to solve this problem, the present inventor cut out the character string to be recognized from the previously input document into a rectangle with width W and height H, and took advantage of the fact that the width of each character is almost constant. proposed a character recognition device that extracts individual character patterns from a character string.

Problems to be Solved by the Invention However, in reality, many characters include ``bias'' and ``structure,'' and the method of cutting out individual characters using only a certain width as a standard causes ``bias'' and ``shape.''``Tsukuri'' was confused with adjacent characters, resulting in errors in cutting out individual characters. The present invention aims to solve the above problems,
To provide a character recognition device capable of correctly cutting out individual characters from a character string and performing character recognition even for characters including ``bias'' and ``tsukuri''.

Means for Solving the Problems In order to solve the above-mentioned problems, the present invention obtains a histogram by scanning a character string cut out in a rectangular shape with a width W and a height H in a direction perpendicular to the character direction. Find the sub-character pattern that is a constituent element of the character pattern by detecting the break in the character string, for example, find the maximum value from the height H of the cut out rectangle and the width _i of the sub-character pattern in the character string. The standard width of the character pattern is A, and the standard width A is
In the process of extracting individual character patterns by combining sub-character patterns based on and determining the width C w _k of the individual character pattern P _k , the sub-character patterns P s _i are extracted from adjacent individual character patterns P _k , When included in P _k+1 at the same time, the individual character pattern P _k ,
The widths C w _k and C w _{k+1 of} P k+1 are compared, and the sub-character pattern P s _i is included in the narrower individual character pattern among the individual character patterns P _k and P _k+1. It consists of

Effect The present invention uses the above-mentioned technical means to accurately cut out even characters that include ``bias'' and ``tsukuri''.
Character recognition becomes possible.

Embodiments Hereinafter, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is 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 including 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 string cutting section 3 perpendicularly to the column direction, obtains a histogram of the character portion, and extracts sub-character patterns which are constituent elements of the character pattern. Reference numeral 5 denotes an individual candidate character pattern extraction section, which extracts individual candidate character patterns from the combination of sub-character patterns extracted by the sub-character pattern extraction section 4. Reference numeral 6 denotes an individual character pattern determination unit, which determines whether a sub-character pattern is included in an adjacent individual character pattern at the same time in each individual character pattern obtained by the individual candidate character pattern extraction unit 5. Uniquely determine the pattern. 7 is a recognition part;
The characteristic amounts such as strokes of each individual character pattern obtained by the individual character pattern determining section 6 are determined and compared with the characteristic amounts of characters registered in advance in the dictionary 8, and the most similar character is selected as a recognition candidate character. A display section 9 displays the recognition results obtained by the recognition section 7.

The character recognition device configured as described above will be explained using an input image shown in FIG. 2 as an example.

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 section 3 cuts out a character string whose absolute position is determined in advance from the input image stored in the image memory 2 in a rectangle R as shown in FIG. 3a.

Next, the sub-character pattern extracting unit 4 scans the character string cut out in the rectangle R perpendicular to the column direction to obtain a histogram of the character string as shown in FIG. 3b. The width of each sub-character pattern w _i
Find (i=1, 2, ..., 8). The sub-character patterns P _s1 , P _s2 ..., cut out in Fig. 3c
Indicates P _s8 .

The individual candidate character pattern extraction unit 5 compares the width w _i of the sub-character pattern from among the sub-character patterns extracted by the sub-character pattern extraction unit and the height H of the character string cut out by the rectangle R, and calculates the maximum Let this value be the reference value A. For example, in FIG. 3b, H is the maximum, and the reference value A becomes H. Furthermore, by combining adjacent sub-character patterns, the sub-character pattern width w _i
If the width between sub-character patterns b _i satisfies the condition |Σ w _i +Σ b _i −A|≦α (α: constant), adjacent sub-character patterns are combined into one individual candidate character pattern, and the individual candidates are Character patterns P ₁ , P ₂ , . . . , P ₇ are obtained as shown in FIG. 4a.

If the sub-character pattern P _si is simultaneously included in adjacent individual character patterns in the individual character pattern obtained by the individual candidate character pattern extraction section, the individual character pattern determining section 6 selects the sub-character pattern P si adjacent to the left and right of the sub-character pattern P _si . The distance between the sub-character patterns is determined, and the sub-character pattern P s _i is included in the individual character pattern to which the closer sub-character pattern belongs. For example, in Figure 4a, the subcharacter pattern
P _s4 is included in the individual candidate character patterns P ₃ and P ₄ at the same time, but since the individual character pattern width C _w3 is narrower than C _w4 , the sub character pattern P _s4 is considered to be a part of the individual candidate character pattern P ₃ , Sub character pattern P _s5
is included in the individual candidate character patterns P ₄ and P ₅ at the same time, but since the individual character pattern width C _w5 is narrower than C _w4 , the sub character pattern P _s5 is considered to be a part of the individual candidate character pattern P ₅ , and the fourth Individual character patterns are determined as shown in Figure b.

The recognition unit 7 determines whether M or more individual character patterns P _i obtained by the individual character pattern determination unit are connected in the direction indicated by the arrow in FIG. 5b, including the pixel of interest, and sets a direction code. , the connectivity of each pixel is examined for each direction code, strokes are extracted, and feature quantities such as the number, position, and length of strokes are extracted. Figure 5a shows the stroke extraction results for the character ``bun''. The extracted feature amount is compared with the feature amount registered in the dictionary 8, and the most similar character is set as a recognition candidate character and displayed on the display unit 9.

Effects of the Invention According to the present invention, when extracting individual character patterns from a character string to be recognized,
It is possible to accurately extract individual character patterns even for character strings containing , and the accuracy of character recognition can be improved.

For example, in Figure 6a, the recognition target character "1"
string' is decomposed into four sub-patterns, P _s10 , P _s11 , P _s12 , and P _s13 , and individual candidate character patterns P ₁₀ , P ₁₁
is obtained as shown in Figure 6b, and the subcharacter pattern
P _s2 is included in both individual character patterns P ₁₀ and P ₁₁ at the same time. In this case, if the distances between sub-character patterns adjacent to P _s2 are not compared, two individual character patterns P ₁₂ and P ₁₃ as shown in Figure 6c will be obtained, resulting in a cutting error, and P _s2 will be It can be seen that individual character patterns P ₁₄ and P ₁₅ as shown in FIG. 6d can be correctly obtained by comparing the distances between adjacent sub-character patterns.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a character recognition device according to an embodiment of the present invention, FIG. 2 is a diagram showing an example of an input image, FIG. 3 is an explanatory diagram of a method for cutting out sub-character patterns from a character string, and FIG. The figure shows the results of cutting out individual character patterns, FIG. 5 is an explanatory diagram of the character recognition method, and FIG. 6 shows the results of cutting out individual character patterns. 1... Image input section, 2... Image memory section, 3...
...Character string extraction section, 4...Sub character pattern extraction section, 5...Individual candidate character pattern extraction section, 6...
...Individual character pattern determination unit, 7...Recognition unit, 8...
...Dictionary, 9...Display section.

Claims (1)

[Scope of Claims] 1. An image input unit that inputs an image including characters to be recognized, and a character string that is a set of characters to be recognized from the image input by the image input unit and has a width w and a height H. Consisting of a character string cutting part cut out in a rectangle, a histogram of pixels forming a character by scanning perpendicularly to the character string direction in the rectangle, and consecutive character parts in which the value of the histogram is 1 or more. a sub _- character pattern extraction unit that extracts a sub-character _pattern that _is If the distance between the character patterns P s _i+1 is d _i , then the width w _i,j between the sub-character pattern P s _i and the sub-character pattern P s _j (=w _i +d _i + _{w i+1} +... …＋w _j )
is less than or equal to the character standard width A, sub-character patterns P s _i , P s _i+1 ..., P s _j are considered as one individual candidate character pattern P _k , and the width w _i,j is an individual character pattern.
An individual candidate character pattern extracting section with a width C w _k of P _k and a sub-character pattern P s _i obtained in the sub-character pattern extraction section are divided into the adjacent individual candidate character patterns P _k and P _k+1 . When included at the same time,
The width C w _k of the individual candidate character pattern P _k , P _k+1 ,
an individual character pattern determining unit that uniquely determines the individual character pattern to which the sub character pattern P s _i belongs by comparing C w _k+1 ; and calculating the characteristics of the character pattern obtained by the individual character pattern determining unit. A character recognition device comprising: a recognition unit that extracts recognition candidate characters by comparing the features with a dictionary.