Kokusai Denshin Denwa Kabushiki Kaisha

Canon Kabushiki Kaisha

Kokusai Denshin Denwa Co., Ltd.

Toshiaki Endoh

Yasuhiro Yamazaki

Tadashi Yoshida

Naoto Kawamura

Hisaharu Kato

Satoshi Katsuno

An original grayscale image or color image to be coded is processed by a bi-level image mingled still image coding system to produce a differential image. The original image and the differential image are compared for each pixel block. Only those of the pixel blocks for which the bi-level image mingled still image coding system will be more efficient than a natural image coding system are encoded by the bi-level image mingled still image coding system, and the other pixel blocks are encoded by the natural image coding system. The coding method of the present invention is directed to a grayscale or color image containing an area where pixel values undergo an abrupt change, such as an edge of a letter, or a document with multi-color and natural color images intermingled therein. The method of the present invention obtains a less deteriorated coded image even from an image in which the edge area is very blunt, and moreover, the invention permits high efficiency coding.

High efficiency coding method for still natural images mingled with

An image reduction system which reduces binary level images is disclosed. In the system, a reduction pixel value is determined by using, as reference pixels for a filter operation: a plurality of surrounding pixels including an object pixel to be reduced; and reduced pixels, from among pixels which have already been reduced, said reduced pixels being at least; a reduced pixel located before the pixel to be determined: a reduced pixel, located on the preceding line, which lies directly above the pixel to be determined; and a reduced pixel before said reduced pixel located on the preceding line. When a pixel referenced during the filter operation is included in a specific pattern, an exception process for correcting binary level results is provided.

Image reduction of binary images using filtering processing

An image reduction system which reduces binary level images is disclosed. In the system, a reduction pixel value is determined by using, as reference pixels for a filter operation: a plurality of surrounding pixels including an object pixel to be reduced; and reduced pixels, from among pixels which have already been reduced, said reduced pixels being at least a reduced pixel located before the pixel to be determined: a reduced pixel, located on the preceding line, which lies directly above the pixel to be determined; and a reduced pixel before said reduced pixel located on the preceding line. When a pixel referenced during the filter operation is included in a specific pattern, an exception process for correcting binary level results is provided.

Image reduction system

A facsimile image coding method in which a pattern appearing only once in the input image is encoded by the conventional facsimile coding method which sequentially scans the input image as a graphic element thereon for encoding, and a pattern appearing twice or more is encoded by encoding the positional coorinates and the identification code added at the time of registration in the pattern library as in the case of the pattern matching method. As a result of the above operations of the present invention, a high compression ratio, not only for a printed document but also for handwritten and graphic documents, is obtained.

Facsimile image encoding method

A graded facsimile image signal coding system is disclosed which is provided with means for distributing a multi-graded facsimile picture signal, represented by 2.sup.n gradation levels, into n bit planes corresponding to n digits; initial encoding means for detecting, from picture elements forming each bit plane, picture elements at intervals of .DELTA.X picture elements on every .DELTA.Y-th line and encoding the detected picture elements; and additional coding means for further coding each one of said encoded picture elements with reference to encoded four picture elements positioned around said one picture element to be coded.

Graded facsimile image signal coding system

A facsimile communication system, in which a coding level of a facsimile signal is determined for each image to be transmitted by a coding algorithm of an encoder in correspondence to an image quality of a receiving image sent from the receiving side of the facsimile signal, so that the above-mentioned each image is encoded in the encoder by the use of the coding level and then transmitted to the receiving side.

Fascimile communication system with desired image quality

A facsimile signal coding system is disclosed, in which initial encoding is performed for detecting, from picture elements forming a facsimile picture signal, picture elements at intervals of .DELTA.X picture elements on every .DELTA.Yth line and encoding the detected picture elements. Encoding of mode 1 is carried out for encoding, by referring to four of the encoded picture elements, a picture element surrounded by the reference picture elements and lying centrally thereof. Encoding of mode 2 is performed by encoding, by referring to four picture elements encoded by the initial encoding and the encoding of mode 1, a picture element surrounded by the four picture elements lying above and below it and on its left and right. The initial encoding and the encoding of mode 1 and mode 2 are effected using 2.sup.n (where n is an integer) as an initial value of each of the .DELTA.X and .DELTA.Y and thereafter the encoding operations of mode 1 and mode 2 are repeated with the values of the .DELTA.X and .DELTA.Y reduced by half.

Growing community of inventors

Tokyo, Japan

Toshiaki Endoh