Kabushiki Kaisha Toshiba

Yoshiaki Kurosawa

Bunpei Irie

Katsumi Kato

Akira Morishita

Satoshi Ito

Yoichi Takebayashi

Mika Fukui

Hisako Tanaka

Seiji Miike

Hideo Horiuchi

Hiroshi Mizoguchi

Yoshikuni Matsumura

Yutaka Hidai

Yasushi Kawakura

Shigenobu Seto

Hideaki Shinchi

Katsunori Ooi

Yoshikatsu Nakamura

In a document image processing apparatus, a document image is inputted as image data from an image inputting section. In a layout analyzing section, the layout of the document image is analyzed on the basis of the image data to obtain layout constituents. In an image processing section, the document image is processed. The image processing section includes an editor for specifying a position to be edited in the document image and editing the document image, on the basis of position/size data on the layout constituents, in accordance with the operator's instructions from the operation data inputting section. In an image displaying section, the document image is displayed in cooperation with the image processing section.

Apparatus and method for processing document image

A document image processing apparatus. A layout analysis section analyzes a layout structure of an input image. A layout information memory section stores layout information representing a relationship between the layout structure and a corresponding area in the input image. An image display section displays the corresponding area in the input image according to the layout information. An indication input section inputs an indication to modify the corresponding area in the input image displayed. A modification section modifies the corresponding area in the input image and the layout structure of the corresponding area in the layout information according to the indication.

Method of analyzing a layout structure of an image using character recognition, and displaying or modifying the layout

An image processing system capable of facilitating the highly accurate character recognition on the colored input images. In this system, an input image data to be processed is entered, and a color separation of the input image data is determined according to geometrical data for each image portion of the input image data, and color data for each image portion and a background of the input image data. Then, the input image data is appropriately processed according to the determined color separation. Also, in this system, at least one of color image data and gray scale image data according to the input image data are stored along with and binary image data according to the input image data, and the binary image data are processed by looking up the at least one of the color image data and the gray scale image data.

Image processing system suitable for colored character recognition

A data retrieval system capable of carrying out the retrieval using the secondary information of the primary data to be retrieved such as environmental information or structural information as the retrieval key. In the system, the primary data which are filing target to be stored and retrieved are stored in a first memory while secondary information associated with the primary data, which are not directly related with data content of the primary data, are stored in a second memory in correspondence to the primary data stored in the first memory. Then, when a retrieval key specifying a desired retrieval of the primary data from the first memory is entered, the secondary information is selected from the second memory in accordance with the retrieval key, and the primary data corresponding to the secondary information selected in accordance with the retrieval key is retrieved and outputted as a retrieval result.

Data retrieval system using secondary information of primary data to be

A pattern recognition system capable of handling the non-Gaussian feature vector distribution in the feature vector space such that the recognition rate can be improved while using the pattern matching based on the assumption of the Gaussian feature vector distribution. In the system, a nonlinear transformation for each recognizable category is applied to the feature vector extracted from the input pattern to be recognized, to obtain a transformed feature data for the input pattern, where the nonlinear transformation maps linearly inseparable distributions in a vector space containing the feature vector onto linearly separable distributions. Then, the transformed feature data for the input pattern is compared with reference feature model for each recognizable category indicating a feature vector distribution for each category, to find a category of the input pattern as that of the reference feature model closest to the transformed feature data for the input pattern.

Pattern recognition system with improved recognition rate using

A pattern recognition apparatus has a contour segmentation unit for dividing an input pattern into segments, a characteristic extraction unit for extracting characteristics of the input segments, and a reference unit for storing characteristic data of reference patterns. The reference unit includes a main reference and a detailed matching reference. The main reference stores partial pattern characteristic data representing the characteristics of segments of each reference pattern. The detailed matching reference stores detailed characteristic data of each reference pattern together with a program for specifying an operation procedures thereof. A matching processor sequentially compares and collates the input pattern with the reference patterns to find out that standard pattern with which the input pattern is matched with the highest similarity. When the input pattern is matched with several reference patterns, a detailed recognition unit performs a detailed recognition process using the detailed characteristic data of these reference patterns to finally select the correct one from among the reference patterns. The main reference additionally stores identification marks to identify specific reference segments necessary to acquired the above detailed characteristic data.

Pattern recognition apparatus

An improved method of decoding in run representation. This method includes the steps of retrieving in response to position information or length information at least one predetermined word pattern containing a plurality of bits arranged in a predetermined bit pattern from a plurality of stored word patterns, each having a different predetermined bit pattern, and processing the retrieved word pattern using OR/Exclusive OR processing to output bit pattern data corresponding to a word unit portion of the original image.

Method of decoding in run representation

A handwritten letter recognition method in accordance with the present invention detects, for each subset, input strokes that coincide with the standard strokes, according to the order of the standard strokes of a standard pattern, and carries out in turn the coincidence processing of strokes by removing the coincident input strokes from the input pattern. The recognition processing for the input patterns is carried out by extracting the standard patterns possessing standard strokes that are coincident for all of the strokes of the input pattern. Therefore, it becomes possible to recognize the input letter pattern easily and efficiently without regard to the order of the strokes input, and even for an input letter pattern with large number of strokes. Moreover, the coincidence detection processing of the strokes can be executed easily and accurately even for the case when the number of strokes that form a letter patterns is large or the case of complicated pattern form, since it detects input strokes that will coincide next, by removing the input strokes that are detected as coincident from the input pattern, to extract a standard pattern all of whose strokes coincide for each and every subset.

Growing community of inventors

Kanagawa-ken, Japan

Yoshiaki Kurosawa