Method and related device for encoding and decoding video images including pixels

Abstract

The present invention relates to encoding decoded video using texture encoding. Specifically, regions of texture in a video image are identified and texture patches are determined for the regions. Clustering is performed to identify regions of texture in the video image. The clustering is further refined. In particular, one or more brightness parameters of a polynomial are determined by fitting the polynomial to the identified texture regions. In the identified texture region, detect pixels whose distance from the fitting polynomial exceeds a first threshold, and use a texture region that does not include one or more of the detected pixels as the corrected texture area. Finally, the corrected texture area is coded separately from the part of the video image that does not belong to the corrected texture area.

Description

technical field [0001] The present invention relates to image and / or video encoding and decoding using texture synthesis. Background technique [0002] Currently, hybrid video codec standards such as H.264 / AVC or H.265 / HEVC employ compression techniques that include predictive coding. Subdivides an image in a video sequence into blocks of pixels and encodes these blocks. The entire block is predicted using the encoded pixels in the spatial or temporal adjacent blocks of a block, instead of encoding the block in units of pixels. The encoder only further processes the difference between this block and its prediction block, usually including transforming the pixels of this block into coefficients in the transform domain. Then, these coefficients are compressed by quantization and entropy coding to generate a code stream. The bitstream also includes any signaling information that enables decoding of the encoded video. For example, the signaling may include encoding settings,...

AI Technical Summary

Problems solved by technology

However, given the nature of the human visual system and the fact that viewers never see raw encoded video, pixel-level fidelity is not a necessary metric

This method of texture synthesis is sufficient to solve the inefficiency of traditional coding methods when dealing with complex textures

Images