Prediction method used for bandwidth compression

Abstract

The invention relates to a prediction method used for bandwidth compression. The method comprises the following steps: acquiring first prediction residual of the current MB according to an adaptive template; acquiring a texture direction gradient value of a plurality of components of the current MB, acquiring a second prediction residual of the current MB; respectively calculating a sum of residual absolute values of the first prediction residual and a sum of residual absolute values of the second prediction residual; selecting the prediction residual of which the sum of the residual absolutevalues is relatively smaller from the first prediction residual and the second prediction residual; and transmitting the final prediction residual of the current MB and mark information of a prediction method corresponding to the final prediction residual of the current MB into a code stream. Compared with the existing method, by a prediction selection algorithm, an optimal prediction method is selected from various prediction methods, when the optimal prediction method is used for processing a complex texture image, the optimal prediction method has excellent prediction effect and high processing efficiency, and can reduce theoretical limit entropy and increase bandwidth compression rate.

Description

technical field [0001] The invention relates to the technical field of compression, in particular to a prediction method for bandwidth compression. Background technique [0002] As people's demand for video quality continues to increase, image resolution, as an important feature of video quality, has transitioned from 720p and 1080p to the current 4K video resolution, and the corresponding video compression standard has also transitioned from H.264 to H. 265. For video processing chips, the multiple increase in resolution will not only cause a substantial increase in chip area cost, but also have a great impact on bus bandwidth and power consumption. [0003] In order to overcome this problem, the bandwidth compression technology applied in the chip is proposed. The goal of on-chip bandwidth compression is to increase the compression factor as much as possible with a smaller logic area cost. On-chip compression is divided into two types: lossy compression and lossless com...

AI Technical Summary

Problems solved by technology

However, when the texture of the image to be compressed is complex and changeable, the correlation is often poor when predicting the complex texture area of ​​the image to be compressed, and the predictive coding cannot get an accurate reference, resulting in the reduction of the theoretical limit entropy. Predicting the quality of the module has become a problem that needs to be solved at present

Images