Fujitsu Frontech Limited

Shohei Hasegawa

Shinichi Eguchi

Maki Yabuki

Katsutoshi Kobara

Hajime Kawashima

Kouichi Kanamoto

Koichi Kanamoto

Yasuyuki Kudoh

Hiroki Inoue

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.

Form recognition apparatus, method, database generation apparatus, method, and storage medium

A gradation distribution of an image character is analyzed, the character color of the image character is analyzed, a gradation ratio between a turning point of a character stroke and the character stroke other than the turning point is analyzed, contrast at an edge of the character stroke is analyzed, the vertical size and center position of the image character are analyzed, the character size of the image character and pitch are analyzed, the first to sixth scores that quantify the respective analysis results are aggregated to discriminate whether the image character is handwritten or printed, and the recognition target item is checked to determine whether it includes both handwritten and printed characters. This makes it possible to enhance accuracy of handwritten-printed character discrimination, which leads to reducing a processing time for character recognition and enhancing accuracy of the character recognition.

Character recognition method, character recognition apparatus, and character recognition program

An objective is to eliminate dotted lines in a character box in image data to increase the character recognition rate. There are some cases in which a dotted line candidate cannot be extracted due to many overlapping parts of dotted lines and characters or due to a blurry part in a dotted line. In such cases, the position of a dotted line candidate is estimated referring to features such as the interval, length, width, etc. of a dotted line candidate in the same character box (or in a character box for another relevant item), and image data of the estimated position and image data of a previously extracted dotted line (or a reference dotted line) are compared to determine whether or not they are an identical dotted line.

Character recognition apparatus and character recognition method

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.

Growing community of inventors

Inagi, Japan

Shohei Hasegawa