Abstract: One system to which the present invention is applied obtains the digitized form image of a form, recognizes a character string existing in the obtained form image, extracts a headline wording being a predetermined character string from the recognized character strings, determines a table structure existing in the form image, on the basis of the extracted headline wording and the arrangement of headline wordings in the form image and specifies a correspondence relationship between a headline wording and a character string other than the headline wording that is recognized, using the determination result.

Abstract: The present invention firstly roughly classifies an analysis range specified by the operator in the color image data of a form into background, a character frame and a character, precisely specifies a character frame on the basis of the classification result, eliminates the character from the color image data from which the background is eliminated and recognizes the remaining character.

Abstract: An edge detecting method detects edge segments by searching all of the search lines forming an image from an end of the image in a direction perpendicular to the edges. If a line whose edge segment cannot be detected exists, a search is made in all of the search lines from the vicinity of the center of the image toward the end of the image, whereby edge segments are detected. A linear edge is determined from edge segments. A plurality of edge candidates are obtained from the edge segments for all of the search lines, and an optimum candidate is selected from among the edge candidates. Ruled lines are extracted from the source document in the image, and an optimum candidate is selected based on a comparison with a ruled line. As a result, an edge of the source document can be detected with high accuracy even if an image on a background side is unstable, or if materials of the background and the source document are similar.

Abstract: It is an object of the present invention to improve the compression ratio of a color image and to clearly display the outlines of characters and the like. A hue cluster classifying/unifying unit reduces the number of hue values of each pixel in a color image, based on a hue histogram, allocates the number-reduced hue value to each pixel and classifies pixels with the same hue value into one cluster. Furthermore, the unit unifies clusters whose hue values are below a predetermined value. The unit also traces the outline of a cluster whose size is below a reference value and determines that a cluster that has a lot of change points belongs to a character area. An encoding unit determines the characteristic of each cluster, based on both an area determined by an area determining unit and whether the cluster belongs to a ruled line area or a character area, and encodes pixels in each cluster by a coding method suitable for the characteristic of the cluster.

Abstract: A method for identifying payment forms accurately identifies types of forms without adding special form identification data. A payment form discrimination method for discriminating payment forms which state a payee account number and a payment amount includes a step of acquiring an image of the form, a step of making a search for the payee account number in the image in accordance with an account number searching rule, and a step of discriminating the type of form based on the searched payee account number. Types of forms can be identified accurately and fast without adding special form identification data, because the form identification is performed using easily searchable account numbers.

Abstract: One system to which the present invention is applied obtains the digitized form image of a form, recognizes a character string existing in the obtained form image, extracts a headline wording being a predetermined character string from the recognized character strings, determines a table structure existing in the form image, on the basis of the extracted headline wording and the arrangement of headline wordings in the form image and specifies a correspondence relationship between a headline wording and a character string other than the headline wording that is recognized, using the determination result.

Abstract: The present invention firstly roughly classifies an analysis range specified by the operator in the color image data of a form into background, a character frame and a character, precisely specifies a character frame on the basis of the classification result, eliminates the character from the color image data from which the background is eliminated and recognizes the remaining character.

Abstract: An image processing device includes a feature emphasis unit for extracting first image frequency information from image data for each first unit area, a boundary provisional determination unit for defining a value obtained by adding a predetermined weight to the first image frequency information as representative feature information, and provisionally determining as a boundary a first unit area whose variance from the representative feature information of the adjacent area is at or higher than a predetermined level, and a boundary determination unit for extracting the second image frequency information for each second unit area smaller than the first unit area in the range of a provisionally determined position and the vicinity, and generating boundary information using as the boundary a second unit area whose value based on the variance from the second image frequency information of the adjacent area is at or higher than a predetermined level.

Abstract: The present invention is a character recognition apparatus, which comprises a background discriminating section, a non-character line discriminating section, a first non-character line removed image creating section that creates a first non-character line removed image, which is an original image from which the non-character line is removed, a first character area discriminating section, an enlarged image creating section, a second non-character line removed image creating section, an interference judgment section that judges whether or not the character and the non-character line interfere with each other in the original image, a character image restoring section that restores the character image, a second character area discriminating section, and a character recognizing section that digitizes the character area recognized by the second character area discriminating section and recognizes the character, thereby characters written on a color form are recognized at a high accuracy.

Abstract: Collation can be carried out at a high speed without reducing the accuracy thereof, whereby processing of identifying documents is accurately performed at a high speed. To this end, characteristic data of a predetermined pattern is registered in advance. Characteristic data in a first area greater than an area of the predetermined pattern registered in advance in an image to be identified is compared and collated with characteristic data of the predetermined pattern. A second area smaller than the first area is cut out from the first area based on the result of comparison, so that characteristic data of an image in the second area is compared and collated with the characteristic data of the predetermined pattern, thus making it possible to identify the predetermined pattern contained in the image based on the result of comparison.

Abstract: An edge detecting method detects edge segments by searching all of the search lines forming an image from an end of the image in a direction perpendicular to the edges. If a line whose edge segment cannot be detected exists, a search is made in all of the search lines from the vicinity of the center of the image toward the end of the image, whereby edge segments are detected. A linear edge is determined from edge segments. A plurality of edge candidates are obtained from the edge segments for all of the search lines, and an optimum candidate is selected from among the edge candidates. Ruled lines are extracted from the source document in the image, and an optimum candidate is selected based on a comparison with a ruled line. As a result, an edge of the source document can be detected with high accuracy even if an image on a background side is unstable, or if materials of the background and the source document are similar.

Abstract: The present invention is a character recognition apparatus, which comprises a background discriminating section, a non-character line discriminating section, a first non-character line removed image creating section that creates a first non-character line removed image, which is an original image from which the non-character line is removed, a first character area discriminating section, an enlarged image creating section, a second non-character line removed image creating section, an interference judgment section that judges whether or not the character and the non-character line interfere with each other in the original image, a character image restoring section that restores the character image, a second character area discriminating section, and a character recognizing section that digitizes the character area recognized by the second character area discriminating section and recognizes the character, thereby characters written on a color form are recognized at a high accuracy.

Abstract: It is an object of the present invention to improve the compression ratio of a color image and to clearly display the outlines of characters and the like. A hue cluster classifying/unifying unit reduces the number of hue values of each pixel in a color image, based on a hue histogram, allocates the number-reduced hue value to each pixel and classifies pixels with the same hue value into one cluster. Furthermore, the unit unifies clusters whose hue values are below a predetermined value. The unit also traces the outline of a cluster whose size is below a reference value and determines that a cluster that has a lot of change points belongs to a character area. An encoding unit determines the characteristic of each cluster, based on both an area determined by an area determining unit and whether the cluster belongs to a ruled line area or a character area, and encodes pixels in each cluster by a coding method suitable for the characteristic of the cluster.

Abstract: An image processing device includes a feature emphasis unit for extracting first image frequency information from image data for each first unit area, a boundary provisional determination unit for defining a value obtained by adding a predetermined weight to the first image frequency information as representative feature information, and provisionally determining as a boundary a first unit area whose variance from the representative feature information of the adjacent area is at or higher than a predetermined level, and a boundary determination unit for extracting the second image frequency information for each second unit area smaller than the first unit area in the range of a provisionally determined position and the vicinity, and generating boundary information using as the boundary a second unit area whose value based on the variance from the second image frequency information of the adjacent area is at or higher than a predetermined level.

Abstract: A data sheet identification device of the invention includes: a character/graphics extracting section, an identical shape deciding section, a graphics collating section, an identification code/data sheet ID identifying section for collating characters that have been decided to have the same shape with an identification code/data sheet ID database in which a plurality of characters showing features of a plurality of data sheets respectively have been registered, and an identifying section for uniquely identifying the data sheet based on a result of the collation by the graphics collating section and a result of the collation by the identification code/data sheet ID identifying section.

Abstract: A method for identifying payment forms accurately identifies types of forms without adding special form identification data. A payment form discrimination method for discriminating payment forms which state a payee account number (50) and a payment amount (51) includes a step of acquiring an image of the form, a step of making a search for the payee account number in the image in accordance with an account number searching rule, and a step of discriminating the type of form based on the searched payee account number. Types of forms can be identified accurately and fast without adding special form identification data, because the form identification is performed using easily searchable account numbers.

Abstract: Disclosed is a format recognition method, apparatus and its storage medium for automatically recognizing the format of a form, whereby the format is automatically determined by examining the arrangement of the smallest rectangles. According to the present invention, the smallest rectangles are extracted from a form, and the positional relationship of these rectangles is obtained. The attribute of the smallest rectangle is determined from the positional relationship. In accordance with the attribute, the smallest rectangles are sorted into a headline portion and a data portion, and a character string in the data portion is recognized.

Abstract: Disclosed are a method of and an apparatus for extracting a dotted line from an binary image of a document, and a storage medium thereof. The isolated points are extracted from the binary image. The isolated points configuring a candidate of the dotted line are extracted based on a positional relationship between the extracted isolated points. A validity of the isolated points configuring the candidate of the dotted line is checked. The dotted line from a positional relationship between groups of the extracted isolated points of the candidate of the dotted line. The dotted line can be thereby precisely extracted even if some isolated points are lost due to an. under-density of the image etc.

Abstract: Data is encrypted and stored in an area on a storage medium accessible by the user outside an external storage device. A decrypting algorithm is stored in an area inaccessible by the user outside the external storage device. The external storage device provided with the storage medium retrieves the decrypting algorithm according to which the data is decrypted using a key obtained from outside the external storage device, for example, from a personal computer connected to the external storage device. Since the encrypted data and its decrypting algorithm are stored on the same storage medium, a specific decrypting algorithm can be assigned to each storage medium, thereby improving the security level for the stored information.

Abstract: Collation can be carried out at a high speed without reducing the accuracy thereof, whereby processing of identifying documents is accurately performed at a high speed. To this end, characteristic data of a predetermined pattern is registered in advance. Characteristic data in a first area greater than an area of the predetermined pattern registered in advance in an image to be identified is compared and collated with characteristic data of the predetermined pattern. A second area smaller than the first area is cut out from the first area based on the result of comparison, so that characteristic data of an image in the second area is compared and collated with the characteristic data of the predetermined pattern, thus making it possible to identify the predetermined pattern contained in the image based on the result of comparison.