Image encoding method, device thereof, and control program thereof

Abstract

A block dead zone scale generator (202) receives an image signal and a prediction error, analyzes the pattern or prediction performance of a target block, and outputs a dead zone scale suitable for the pattern or prediction performance of the block. A dead zone generator (201) receives a dead zone scale from the block dead zone scale generator (202) and an MB quantization parameter from a quantization control device (103), calculates a dead zone width from the zone scale and the MB quantization parameter, and outputs the dead zone width. A quantization device (102) quantizes an orthogonal transformation coefficient supplied from an orthogonal transformation device (101) by using a dead zone from the dead zone generator (201), and outputs a quantized transformation coefficient. This makes it possible to realize quantization with arbitrary strength for each transformation coefficient and for each block comprising a plurality of transformation coefficients as constituent elements, thereby providing a high-quality image encoding technology.

Description

technical field [0001] The present invention relates to an image encoding method, an apparatus thereof, and a control program thereof, and particularly to an image encoding method performing adaptive quantization for improving subjective image quality, an apparatus thereof, and a control program thereof. Background technique [0002] The prior art will be described below with reference to the accompanying drawings. [0003] First, hybrid moving image coding as a prior art based on an orthogonal transform device and a prediction (intra prediction / inter prediction) device will be described with reference to FIG. 1 . [0004] According to the related art, an image frame forming a moving image is divided into a plurality of regions called macroblocks (MBs), and each block obtained by further dividing each MB is encoded. FIG. 2 shows the structure of an AVC (Advanced Video Coding: ISO / IEC 14496-10) image frame as a specific example of the image frame structure according to the p...

AI Technical Summary

Problems solved by technology

[0017] The first problem is that the individual blocks that make up the MB do not necessarily have the same pattern

In this case, the prior art cannot perform quantization suitable for the pattern of each block constituting the MB

[0018] The second problem is the performance of minimizing the prediction error (hereinafter called predictive performance) differs for each block that makes up the MB

In this case, the prior art cannot perform quantization suitable for the predictive performance of each block composing the MB

[0019] The third problem is that the distribution of the orthogonal transform coefficients corresponding to the coordinates in the block (hereinafter referred to as spatial frequency) varies due to the first and second reasons, and the individual blocks constituting the MB do not show a uniform distributed

In this case, the prior art cannot perform quantization suitable for the distribution of orthogonal transform coefficients for each block

Images