Image encoding method and image encoder, image decoding method and image decoder, and image processing system

Abstract

A method and an apparatus improving the coding efficiency by representing encoded data in a compact form by means of a pattern matching technique. A waveform vector similar to a partial signal waveform of an image signal in a predetermined group of waveform vectors is searched for. When a compression encoding is performed for information to identify the waveform vector searched for, similarity information, and a position in an image signal of the partial signal waveform in accordance with a predetermined encoding rule, the position in the image signal of the partial signal waveform is encoded through replacement with position information in a predetermined partial domain of the image signal. As a result; since the position in the image of the partial signal waveform that is the most similar to the waveform vector is encoded through replacement with the position information in the predetermined partial domain of the image signal, it is possible to represent the position more compactly.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS [0001] This application is a Divisional of U.S. patent application Ser. No. 10 / 297,488, filed Dec. 9, 2002, which is a National Stage of International Application PCT / JP02 / 03498, filed Apr. 8, 2002, and claims priority to Japanese Patent Application No. 2001-110663, filed Apr. 9, 2001.BACKGROUND OF THE INVENTION [0002] The present invention relates to an image encoding and decoding method for encoding a source image such as a moving image and a still image and decoding the encoded image and, more particularly, to an image encoding method, an image encoder, an image decoding method and an image decoder that can encode an image and decode the encoded image with no use of DCT (discrete cosine transform) employed in standard algorithms of MPEG series and so on. [0003] Furthermore, the present invention relates to an image processing system that encodes and decodes an image in accordance with such an image encoding method and an image decoding meth...

AI Technical Summary

Benefits of technology

[0011] It is a first object of the present invention to provide a method and an apparatus that improve total coding efficiency by compactly representing encoded data by means of the pattern matching technique.

[0018] When the number of points at which it is determined whether or not the partial waveform signal and the waveform vector match are decreased, it becomes possible to reduce the number of operations. From the viewpoint, as claimed in claim 3, in the above-mentioned pattern matching step, pixels formed of a predetermined power of 2×2 of an image signal represented by a predetermined pixel unit may be set as a unit, and a position in an image of a partial signal waveform to be encoded may be identified.

[0024] According to such an image encoding method, a group of waveform vectors to be used is selected among a plurality of groups of waveform vectors based upon the parameter used for performing an encoding at the first image encoding step and the position in the second image of the partial waveform signal. As a result, it is possible to select a group of waveform vectors corresponding to a characteristic of each image to be encoded and to encode the image signal more appropriately.

[0025] The second image information is a signal corresponding to an error signal between a first image signal provided to the first encoding step and a signal obtained by performing a local decoding process for encoding information obtained at the first image encoding step. As a result, when the first image information provided to the first image encoding step is assumed to be a source image, the second image information corresponds to an encoding distortion signal generated by encoding the source image. In such a case, according to the image encoding method, it is possible to efficiently encode a signal corresponding to the error signal ending up with unnecessary additional information when the source image is encoded.

Problems solved by technology

On the other hand, when encoding is performed at an extremely low bit rate, a resulting coarse quantization degrades a reconstruction of the coefficients.

Consequently, there arise some problems in that it is impossible to reconstruct important bases to an intrinsic signal representation.

Also, since the DCT operates a closed process on an 8×8 image block, the DCT has the tendency that a distortion caused by quantization noticeably appears in a boundary of blocks.

That generates a block distortion and exhibits in the image an element that the original signal does not contain visually, whereby the element is recognized as a seriously noticeable noise.

As a result, the coefficient in the high frequency, which plays an important role in reconstructing the waveform, is lost.

The loss of the coefficient causes harmful noise peculiar to the DCT and results in image quality degradation.

In order to take advantage of the “Matching Pursuits” encoding technique, however, the problem that there is a burden on implementation such as the number of operations necessary for the encoding side to perform the basis search is pointed out.

As a result, if little information is used in the enhance layer, the encoding does not work efficiently.

Images