Methods and apparatus for video coding

Abstract

A method and apparatus of video coding using position dependent matrix weighted angular intra prediction are disclosed. According to this method, at the decoder side, a reference template area is determined for the current block. A storage-reduced set of matrix weights is received. The storage-reduced set of matrix weights is converted into a set of matrix weights by applying a target inverse storage-reduction process to the storage-reduced set of matrix weights, wherein the target inverse storage-reduction process includes at least one of: unmasking coefficients of the storage-reduced set of matrix weights, up-sampling the coefficients of the storage-reduced set of matrix weights, applying image decompression to the coefficients of the storage-reduced set of matrix weights, or dequantizing the coefficients of the storage-reduced set of matrix weights. An intra prediction for the current block is generated using the set of matrix weights and the reference template area.

Images