Apple Inc.

Other

Xiang Fu

Xiaohua Yang

Linfeng Guo

Francesco Iacopino

Felix Chou

Hao Pan

Haitao Guo

Wei Li

Krishnakanth Rapaka

Ying Jian He

Haiyan He

Munehiro Nakazato

Xu Zhao

Krishna Rapaka

Mukta Gore

Techniques for encoding video with reduced frame-boundary artifacts are presented. The techniques include a video encoding method, where, when skip mode is selected as a motion prediction coding mode for a pixel block predicted from a reference block of a reference frame, estimating an amount of the reference block that extends beyond the edge of the reference frame. If the amount beyond the edge does not exceed a threshold, the prediction may be coded with skip mode. If the amount beyond the edge exceeds a threshold, the pixel block may be partitioned, and the partitioned blocks may be coded with motion prediction.

Frame boundary artifacts removal

Techniques are disclosed for coding image data adaptively at different levels of downscaling. Such techniques may involve partitioning input data into pixel blocks for coding and performing content analysis on the pixel blocks. The pixel blocks may be input to block coders that operate at different pixel block sizes, which may code the pixel blocks input to them at their respective sizes. Except when a block coder operates at the partitioning size, block coders that operate at different pixel block sizes may perform downscaling of the pixel blocks to match their size with the block coders' respective coding size. A block decoder may invert the coding operations performed by the block coders, decoding coded image data at respective pixel block sizes, then upscaling decoded image data obtained therefrom to a common pixel block size. Image reconstruction may synthesize a resultant image from the decode pixel block data output by the decoders. Quality enhancements may be added by neural network processing systems that have been trained to other segmented images.

Deep quality enhancement of adaptive downscaled coding for image compression

Techniques are disclosed for developing quantization matrices for use in video coding. According to these techniques a first quantization matrix may be derived from a second quantization matrix by scaling quantization values of the second quantization matrix by scaling parameters. The scaling parameters may increase according to distance between each matrix position and a matrix origin, they may be derived from characteristics of a video sequence to be coded, or both. The first quantization matrix may be communicated to a decoder. Thereafter, a video sequence may be coded predictively. As part of the coding, pixel data of the video sequence may be transformed to a plurality of frequency domain coefficients, and the frequency domain coefficients may be quantized according to the first quantization matrix.

Techniques for calculation of quantization matrices in video coding

Growing community of inventors

Mountain View, CA, United States of America

Xiang Fu