JPH01279385A - Character recognizing device - Google Patents

Abstract

PURPOSE:To recognize an inverted character which is written in white on black ground through an ordinary operation by making both normal character pictures written in black on white paper and inverted character pictures written in white on black paper recognizable. CONSTITUTION:A picture inverting means 7 is provided and the number of white picture elements and black ones of inputted pictures 12 and 13 are calculated. When an inputted picture is black characters written on white ground, the white picture element predominates the black picture element in number and, when white characters written on black ground, the black picture element predominates the white picture element is number, in most case. Therefore, by comparing the number of white picture elements with that of the black picture elements and discriminating whether an inputted picture is the inputted picture 12 or 13, the inputted picture is not processed further and treated as an inversion-processed picture 15 when the inputted picture is the picture 12. When the inputted picture is the picture 13, the white and black picture elements are inverted to each other and the inputted picture is treated as the inversion-processed picture 15. Therefore, both black characters written on white ground and white characters written on black ground can be recognized by means of an ordinary recognizing means.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field 1] The present invention relates to a character recognition device that searches for a character area from a character image and converts it into a code number by manually inputting characters written on a sheet of paper as an image. .

[Prior art 1] A character recognition device inputs characters written on paper as an image into a storage device using an input device such as a scanner, automatically searches for the character position from the captured image data, and processes the search result. The character position data thus obtained is compared with character data in a dictionary to determine what kind of character the character image is, and the character code is converted into a character code and output. Generally, when searching for a character position from a human character image, the following procedure is used.6 Normally, a human character image is input in multiple lines at the same time. Therefore, each row must be cut out from the input image. At this time, the peripheral distribution in the row direction is taken, and a cluster of black pixels is determined to be a character line based on the peripheral distribution, and the line is cut out. Next, the peripheral distribution in the direction perpendicular to the row direction is taken, and similarly to the row direction, a cluster of black pixels is determined to be a character position, and the character position is determined. After that, the data of the determined character position (for example, peripheral characteristics Journal of the Institute of Electronics and Communication Engineers '82 Vol. J65-D No. 8P
026-1033, Metzhu characteristic research practical application report No. 34
Characters are recognized by comparing the character data held in the dictionary (Volume No. 1, P, P, 47-58) and finding the closest one, and all the processes from character extraction methods to recognition are performed. The method targets ordinary character images such as black characters on a white background.

[Problem to be solved by the invention 1 However, the character images that are usually used are not only black characters on a white background, but also quite a lot of bright characters such as white on a black background or a dark background close to black. When a character image is input, it is impossible to cut out and recognize the characters. In addition, when trying to recognize black and white inverted characters, character extraction must be performed by determining that a white block is a character area, and then comparing it with character image data created for inverted characters. The amount of dictionaries will be doubled. Further, to determine whether or not a character is an inverted character, it is necessary to input it from the outside by some method, and recognizing inverted characters using such a method is extremely inefficient and wastes memory. Therefore, the present invention is intended to solve such problems, and its purpose is to make it possible to recognize not only normal characters such as black characters on a white background, but also black and white inverted characters such as white characters on a black background by normal operations. An object of the present invention is to provide a recognition means and a recognition method.

[Means for Solving the Problems] The present invention provides: (1) an image input means for inputting a document image, a character extraction means for extracting a character string from the input image, and a character extraction means for extracting characters from the extracted character string. In a character recognition device comprising a character recognition means for recognizing the character, and a recognized character display means for displaying the recognition result, for example, an ordinary character image such as a black character written on a white paper, a character image written on a black paper, It is characterized by being able to recognize both inverted character images such as written white characters.

(2) In the character recognition device, if the read image is white characters on a black background, the recognition operation is performed after the read data is inverted in black and white.

(3) Determination as to whether the text is black on a white background or white on a black background is characterized by using statistics of white pixels and black pixels.

[Example] The present invention will be described in detail below based on examples.

As shown in the block diagram of FIG. 1, the character recognition device of the present invention includes a CPU 1. Image human power device 2, recognized character display device 3,
It is composed of ROM4 and RAM5. Further, the operation of the character recognition device of the present invention is as shown in the block diagram of FIG. 2. The image input means 6 determines whether the input image is black characters on a white background or white characters on a black background. Image reversing means 7°- for reversing the image in black and white in the case of white characters on a black background.
character string extraction means 8 for extracting a character string from a character string, character extraction means 9 for extracting characters from a character string, character recognition means lO1 for recognizing the extracted characters; recognition result display means 11 for displaying the recognized characters; It consists of

Hereinafter, the operation of the character recognition apparatus of the present invention will be explained based on the flowchart shown in FIG. 4, regardless of whether the human character image is black characters on a white background or white characters on a black background.

In the image input means 6, the image input device 2
A character image is read into RAM5. Is the data read into the RAM 5 data 12 with black characters on a white background?
If it is not known whether data B is a white character on a black background, and the character string extraction means 8- tries to extract the character, the peripheral distribution in the row direction will be calculated and the black pixel area will be picked up. Here, the present invention provides an image reversing means 7 (reversing processing means 16).

In the reversal processing means 16, the human image 12 (13)
Calculate the number of white pixels and the number of black pixels in the input image.If the input image is black text on a white background, the number of white pixels will be greater than the number of black pixels in most cases, and if the input image is white text on a black background, the number of white pixels will be greater than the number of black pixels. In this case, the number of black pixels is greater than the number of white pixels. Therefore, in FIG. 3 14, the number of white pixels and the number of black pixels are compared to determine whether the input image is 12 or 13.
If the human image is 12, the inverted image is set to 15 without doing anything in 14-1, and if the input image is 13, it is 15.
In step 4-2, white pixels and black pixels are inverted to form an inverted image 15. Since the image 15 obtained in this way has black characters on a white background, the subsequent recognition processing can be performed by normal recognition processing, and the character recognition means 10 needs to newly store character data for white characters on a black background in the ROM 4. This increases the amount of characters that can be recognized per unit of memory. In addition, in the present invention, it is possible to determine whether the text is black on a white background or white on a black background by calculating the number of white pixels and black pixels in all input images, Even if you make a judgment by calculating the number of pixels in only the area, there is almost no misjudgment.6 Calculate the number of white pixels and black pixels of the human image as described above, and determine whether it is an image with black text on a white background or white text on a black background. After determining whether the image is an image with white text on a black background, the white pixels and black pixels of the image are inverted to create a character image with black text on a white background, so normal recognition cannot recognize black and white inverted characters. It becomes possible.

Furthermore, since the reversal judgment is automatically made, the operation is no different from conventional character recognition devices, and only the functionality can be improved.

Furthermore, although this embodiment has described the case where the input image is binary (black and white), this method is not limited to binary images, and even data with gradation can be used to calculate the statistics of i-situation. Take it,
If the character is an inverted character, it is possible to provide a character recognition device with a similar function by inverting the gradation (taking two's complement).

[Effect of the Invention 1] As described above, according to the present invention, a character recognition device automatically determines whether an input image is a black character on a white background or a white character on a black background by taking pixel statistics. In the case of white text, black and white is inverted, making it possible to recognize white text on a black background and black text on a white background using normal recognition means. Moreover, such a character recognition device has a wider range of uses than conventional ones, and is more convenient.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a character recognition device of the present invention. FIG. 2 is a block diagram of the recognition means of the present invention. FIG. 3 is a diagram showing the inversion of the characters of the present invention. FIG. 4 shows a flow chart of character reversal determination according to the present invention. 1... CPU 2... Image input device 3... Recognition result display device 4... ROM 5... RAM 6... Image input means 7 ... Image inversion means, 8 ... Character string extraction means 9 ... Character extraction means 10 ... Character recognition means 11 ... Recognized character display means 12.13 ... Input image 14 ... Reversal processing 15 ... Image after reversal processing 16 ... Reversal processing means and above Applicant Seiko Epson Co., Ltd. Agent Patent attorney Masayoshi Kamiyanagi (1 other person) )Figure 2

Claims (3)

[Claims] (1) An image input means for inputting a document image, a character string extraction means for extracting a character string from the input image, a character extraction means for extracting characters from the extracted character string, and a character recognition means for recognizing the characters. , and a recognized character display means for displaying the recognition results, for example, normal character images such as black characters written on white paper, and inverted character images such as white characters written on black paper. A character recognition device characterized by being able to recognize either. (2) The character recognition means according to claim 1, wherein in the character recognition device, when the read image is white characters on a black background, the read data is inverted in black and white and then the recognition operation is performed. (3) The character recognition means according to claim 1, wherein statistics of white pixels and black pixels are used to determine whether it is a black character on a white background or a white character on a black background.