JPH0654184A - Image recognizing device - Google Patents

Abstract

PURPOSE:To improve the accuracy of recognition, to attain the quick recognition of widespread originals and to realize the prevention of an illegal copying by recognizing an official name, etc., printed on paper money, a valuable securities, etc. CONSTITUTION:This device identifies and recognizes a special original by being provided with a digital color input part 101 inputting a digital image, a specified character string storage part 102 storing a specified character string or (and) a specified numeric string and a discrimination part 103 discriminating whether the specified character string or (and) the specified numeric string stored in the specified character string storage part 102 exists or not in the digital image inputted by the digital color image input part 101.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image recognition apparatus for recognizing whether an input digital image is a special document such as a bill, and more specifically, a specific character storing the input image in advance. The present invention relates to an image recognition device that matches a row or (and) a specific numeral row, determines the degree of similarity, and recognizes whether the read original is a special original such as a bill.

[0002]

2. Description of the Related Art As a conventional image processing apparatus for preventing forgery of banknotes, detection means for detecting the position state of a document, and image data of a stamp portion in the document according to the detection of the detection means are taken out and registered in advance. It has been proposed to provide a discriminating means for discriminating the degree of similarity with the formed pattern and discriminate a banknote based on the degree of similarity of the stamp portion.

Further, as another conventional technique, Japanese Patent Laid-Open No.
There is an "image processing device" disclosed in Japanese Unexamined Patent Publication No. 0-229572. In this image processing apparatus, the image data itself of the banknote is handled in order to specify the content of the original document, and whether the banknote is a banknote or not is determined in correlation with the input data.
Further, there is an "image processing device" disclosed in Japanese Patent Application Laid-Open No. 2-210591. In this image processing apparatus, the empty portion of the banknote is targeted as the information for specifying the content of the document.

[0004]

However, in the above-mentioned prior art, in order to discriminate a bill, the stamp portion on the bill, the image data of the bill itself, and the empty portion of the bill are identified. With regard to emptying, there is a high possibility that similar images will exist, and further, when seals are the object of comparison, there is a problem that official documents without seals cannot be recognized. Further, regarding the image data itself of the banknote, there is a problem that the recognition efficiency is low because the data amount is large and the recognition process cannot be executed quickly.

The present invention has been made in view of the above, and improves recognition accuracy by recognizing official names and the like written on banknotes, securities, etc. The purpose is to enable quick recognition and prevent illegal copying.

[0006]

In order to achieve the above-mentioned object, the present invention comprises an input means for inputting a digital image,
Storage means for storing a specific character string or / and a specific number string, and a specific character string or / and a specific character string stored in the storage means in the digital image input by the input means.
(EN) An image recognition device provided with a discriminating means for discriminating whether or not a specific numeral string exists.

Further, input means for inputting a digital image, storage means for storing a specific character string or (and) a specific number string, character type or (and) number type, and a digital image input by the input means An image recognition apparatus is provided, which includes a determination unit that determines whether or not a specific character string or / and a specific number string stored in the storage unit exists.

Further, input means for inputting a digital image, storage means for storing a specific character string or (and) a specific number string, character size or (and) number size, and a digital image input by the input means An image recognition apparatus is provided, which includes a determination unit that determines whether or not a specific character string or / and a specific number string stored in the storage unit exists.

Further, input means for inputting a digital image, storage means for storing a specific character string or (and) a specific number string, a hatch pattern of characters or (and) a hatch pattern of numbers, and the input means. Another object of the present invention is to provide an image recognizing device including a determination unit that determines whether or not a specific character string or / and a specific number string stored in the storage unit is present in the digital image.

Further, a detecting means for detecting the position state of the original, a recognizing means for recognizing the numerical information on the original based on the detection result of the detecting means, a storage means for storing the numerical information of the special original, (EN) An image recognition device provided with a discriminating means for discriminating whether or not an original is a special original by comparing the recognition information recognized by the recognizing means with the stored information stored in the storage means. .

[0011]

The image recognition apparatus according to the present invention inputs a digital image and, on the digital image, for example, a specific character string or / and a specific numeral string stored in a memory,
The special original is identified by determining whether or not there is a hatch pattern of character type or (and) number type, character size or (and) number size, character or (and) number.

[0012]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the accompanying drawings. FIG. 1 is a block diagram showing a schematic configuration of a document recognition apparatus according to the present invention. In the figure, 101 is a digital color image input unit configured by a color scanner, for example, for reading image information of banknotes, R, G,
B Input an 8-bit digital color image for each color. 10
A specific character string storage unit 2 stores various character strings of the Ministry of Finance Printing Bureau. In particular, the character string is a character string peculiar to banknotes, and since these character strings are also in the background portion in the image information, the cutting process is easy. 10
Reference numeral 3 denotes a determination unit, which is used for the specific character string storage unit 10 for the character string cut out from the input digital image information.
The matching process with the specific character string stored in 2 is executed.

Next, this operation will be described based on the flow chart shown in FIG. First, a document (for example, a bill) is read by the digital color image input unit 101,
8-bit digital color image data for each of R, G, and B is input to the determination unit 103 (S201). Next, the entire character string portion is cut out from the input digital image data (S202). As this method, for example, a dot string (character) existing in the digital image data is selected, and the character string is extracted from the connectivity with the surrounding characters. After that, the specific character string is read from the specific character string storage unit 102 (S203), the extracted character string is matched with the read character string (S204), the similarity is determined, and the determination result is obtained. Recognizes that the read image is a specific document, for example, a bill (S20).
5). A public name can be determined for the character string on the digital image information input by the determination unit 103, and public documents (banknotes, securities) can be selected.

In the above embodiment, the specific character string storage unit 102 for storing the specific character string has been described as an example. However, in addition to the specific character string, a specific numeral string, a character / numeral type, a character / character
Similar configurations and operations can be realized for information such as number size and character / number hatch patterns, and if the amount of information for comparison stored in the (specific character string) storage unit 102 is increased, the determination unit 103 can recognize the information. The accuracy of judgment can be improved. Also, the above-mentioned respective elements may be combined, and by combining them, the recognition rate will be further improved. In particular, the character size (the number of points) and the character type show unique characteristics in an official document, which dramatically improves the recognition rate.

Next, another embodiment according to the present invention will be described. FIG. 3 is a block diagram showing a schematic configuration of the image recognition apparatus according to this embodiment. In the figure, 301 is a manuscript,
Reference numeral 302 denotes an image input device that reads image information from the original 301 and converts it into digital image information. Reference numeral 303 denotes an original position detection unit that detects the position of the original 301 based on the digital image information input by the image input device 302, 304.
Is a character image detection unit that detects a character image from digital image information based on the position information detected by the document position detection unit 303, and 305 is a character recognition unit that recognizes characters in the character image detected by the character image detection unit 304. Thirty
Reference numerals 5 and 306 denote a character image detection unit 304 and a character recognition unit 305.
An information signal from the document is input to match with information read from the specific document information dictionary 307 to determine whether or not there is a match. A determination unit 307 stores various information on the specific document in advance. A manuscript information dictionary, 308 is a determination unit 3 for the image information input from the image input device 302.
The image generation unit includes a series of image generation processes that perform image generation processing based on the information signal from 06.

Next, the operation will be described with reference to the flowchart shown in FIG. The image information of the document 301 is obtained by the image input device 302 (S401), and the document position is detected by the document position detection unit 303 (S402).
Next, the character image detection unit 304 separates the image of the character portion from the image of the original portion (S403).
Features such as size and position information are extracted (S404),
The information is sent to the determination unit 306. Further, the character image part is converted into code information by the character recognition part 305 (character recognition).
The information is also input to the determination unit 306 (S405).

In the judging section 306, the information stored in the specific manuscript information dictionary is read and compared with the information inputted from the character image detecting section 304 / character recognizing section 305 to determine whether the copying is illegal. It is determined whether or not (S40
6). As a result, if it is determined that the copy is not illegal, image output is executed via the image generation unit 308 (S407). Conversely, if it is determined that the copy is illegal, the image is output. Or processing is performed on the image via the image generation unit 308 and then output (S4).
08).

Further, in the character image detecting section 304, the image data of the character part is extracted by utilizing the pictographic character separation technique utilizing the signal level. Further, the character recognition unit 305 processes the image data to obtain character code data. Here, in the character image detection unit 304, 100
% It is difficult to extract only the pixels that completely form the character portion. Therefore, it is difficult for the character recognition unit 305 to completely extract only numbers such as banknotes and other character codes.

Therefore, the specific manuscript information is used by using the character image information (color, size, shape of the pixel cluster of each character corresponding part, positional relationship between characters, position on the manuscript) obtained by the character image detecting unit 304. Compare with information such as banknotes in the dictionary 307,
It is determined whether or not the input document is prohibited from being copied. Further, based on the result of character recognition of only the character portion image corresponding to the position where the issue number of the bill or the like should exist, it is also used for the determination as to whether or not the number code is peculiar to the bill.

The determination unit 306 for determining illegal copying uses the characteristics of the character portion image and the character recognition result. In the character recognition unit 305, by comparing the document position information with the existence position of the character part such as a number in a bill, etc., and performing character recognition only on the pixel at the position where the character part such as the number should be present, it is unnecessary. Prevents the generation of character code information,
It is also possible to suppress erroneous determination in the determination unit 306.
As described above, since both the image information of the character portion in the document and the code information which is the recognition result thereof are used, illegal copy detection can be realized with high accuracy.

[0021]

As described above, according to the image recognition apparatus of the present invention, the input means for inputting the digital image, the storage means for storing the specific character string or / and the specific numeral string, and the input means. Since the digital image inputted by means of the judgment means for judging whether or not the specific character string or / and the specific numeral string stored in the storage means is present, By recognizing the public name and the like, the recognition accuracy can be improved, a wide range of originals can be recognized quickly, and illegal copying can be prevented.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of an image recognition apparatus according to the present invention. It

FIG. 2 is a flowchart showing an operation of the image recognition device shown in FIG. .

FIG. 3 is a block diagram showing a schematic configuration of another image recognition apparatus according to the present invention.

FIG. 4 is a flowchart showing an operation of the image recognition apparatus shown in FIG.

[Explanation of symbols]

101 digital color image input unit 102
Specific character string storage unit 103 Judgment unit 301
Document 302 Image input device 303
Document position detection unit 304 Character image detection unit 305
Character recognition unit 306 Judgment unit 307
Specific manuscript information dictionary 308 Image generation unit

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Takashi Saito 1-3-6 Nakamagome, Ota-ku, Tokyo In-house Ricoh Co., Ltd. (72) Takatsugu Saito 1-3-6 Nakamagome, Ota-ku, Tokyo In Ricoh Co., Ltd. (72) Inventor Shigeo Kurotaka 1-3-6 Nakamagome, Ota-ku, Tokyo Inside Ricoh Co., Ltd.

Claims (5)

[Claims] 1. An input means for inputting a digital image, a storage means for storing a specific character string or / and a specific numeral sequence, and a specification stored in the storage means for the digital image input by the input means. An image recognition apparatus comprising: a determination unit that determines whether or not a character string or / and a specific numeral string exists. 2. An input means for inputting a digital image, a storage means for storing a specific character string or (and) a specific number string, a character type or (and) a number type, and a digital image input by the input means. An image recognition apparatus comprising: a determination unit that determines whether or not a specific character string or / and a specific number string stored in the storage unit exists. 3. Input means for inputting a digital image, storage means for storing a specific character string or (and) a specific numeral string, character size or (and) numerical size, and a digital image input by the input means. An image recognition apparatus comprising: a determination unit that determines whether or not a specific character string or / and a specific number string stored in the storage unit exists. 4. Input means for inputting a digital image, storage means for storing a specific character string or (and) a specific numeral string, a hatch pattern of characters or (and) a hatch pattern of numbers, and the input means. An image recognition apparatus, comprising: a determining unit that determines whether or not a specific character string or / and a specific number string stored in the storage unit exists in the digital image. 5. A detection means for detecting the position state of the document,
Recognizing means for recognizing the numerical information on the original based on the detection result of the detecting means, storage means for storing the numerical information of the special original, recognition information recognized by the recognizing means and the storage means. An image recognition apparatus, comprising: a determination unit that compares stored information with a document to determine whether the document is a special document.