Alibaba Group Holding Limited

Jian Lou

Yu-Chen Sun

Minghai Qin

Yuhao Wang

Yen-Kuang Chen

Yizin Du

Tsui-shan Chang

Systems and methods are provided for implementing inter-coded resolution-adaptive video coding supported by resizing motion information of a current frame during motion prediction, reducing extensive computational costs and computing time that would be incurred as a result of extensive application of interpolation filters to generate picture information of a reference frame to sub-pixel granularity, and thus reducing computational cost and computing time of decoding as a whole. The methods and systems described herein include determining a ratio of the resolution of a reference frame to the resolution of a current frame; determining a motion vector of the block of the current frame, and calculating at least one pixel coordinate corresponding to the motion vector of the block of the current frame; resizing motion information of the block of the current frame according to a resolution of the reference frame in accordance with the ratio; locating a predictor block of the reference frame in accordance with the resized motion information; and subsequently minimizing or avoiding applying an interpolation filter to a block at the scaled coordinates at the reference frame to generate sub-pixel values of the predictor block.

Resolution-adaptive video coding

The present disclosure relates to a computer-implemented method for processing video data. The method comprises receiving a user input corresponding to a first picture of the video data, generating, based on the user input, prediction information of the first picture with respect a reference picture of the video data, and encoding the first picture using the prediction information.

Video encoding technique utilizing user guided information in cloud environment

Systems and methods are provided for improving accuracy and efficiency of a context-adaptive binary arithmetic coding (CABAC) by adaptively selecting a context model specific to the characteristics of a coding unit (CU), such as the size, dimension (height and/or width), type (luma or chroma), and/or flag type (cu_palette_flag or pred_mode_flag) of the CU, that comprise: determining a characteristic of the CU; determining whether the characteristic of the CU meets a corresponding threshold; and upon determining that the characteristic of the CU meets the corresponding threshold, selecting a first context model, or upon determining that the characteristic of the CU fails to meet the corresponding threshold, selecting a second context model.

Context model selection based on coding unit characteristics

Systems and methods are provided for improving efficiency of an encoder performance in the HEVC palette mode by limiting the number of colors to be used in the current palette to be no more than the size of coding unit or the prediction unit, and by avoiding duplicate colors in the current palette. The encoder may signal a binary vector indicating previous elements from a palette predictor list to be reused in a current palette, determine a number of the previous elements based on the binary vector, determine a sample size of the current palette, determine a maximum number of new elements to be used in the current palette based on the number of the previous elements and the sample size, determine a set of new elements corresponding to the maximum number of new elements, and signal the set of new elements.

Palette size constraint in palette mode for video compression system

A device is disclosed. The device includes: an encoder configured to encode a first bit stream and a second bit stream, the first bit stream comprising a first plurality of frames, the second bit stream comprising a second plurality of frames; a processor configured to determine a first encoding parameter of the first bit stream and a second encoding parameter of the second bit stream, the first encoding parameter different from the second encoding parameter; and a communication module configured to transmit the first bit stream and the second bit stream over a network. The first plurality of frames and the second plurality of frames are transmitted in an interleaving fashion.

Bit stream management in video communication

Systems and methods are provided for improving filtering performance and Bj&#xf8;ntegaard-Delta (BD) rate savings for video processing. In addition to computing the artifacts between a given compressed image and a restored clean image after filtering using Deep Residual Learning (DRL) for recovering the residual between input and output, filtering strength of a loop filter may be controlled by the content of the region of the image, such that, in more important areas, such as the face and edges, the filtering strength may be increased while in less important areas, such as textures and backgrounds, the filtering strength may be decreased.

Content-weighted deep residual learning for video in-loop filtering

Systems and methods are provided for implementing methods for resolution-adaptive video coding in a motion prediction coding format by obtaining a current frame of a bitstream, obtaining one or more reference pictures from a reference frame buffer, up-sampling or down-sampling the obtained one or more reference pictures that have resolutions different from a resolution of the current frame, and generating a reconstructed frame from the current frame based on the one or more reference pictures and a selected motion candidate of the current frame, thereby achieving substantial reduction of network transport costs in video coding and delivery without requiring the transport of additional data that would offset or compromise these savings.

Systems and methods are provided for selecting and applying an appropriate deblocking filter for improved coding efficiency and compression performance by partitioning a frame into a plurality of rectangular coding units (CUs) according to the QTBT frame partitioning scheme, selecting a first and a second CUs of the plurality of rectangular CUs, where the second CU is adjacent to the first CU, identifying a CU boundary shared by the first and second CUs, determining that at least one of the first or the second CU is non-square in shape, determining a first shape of the first CU and a second shape of the second CU, adaptively selecting a deblocking filter based on the first shape and the second shape, and applying the selected deblocking filter to pixels across the CU boundary.

Adaptive deblocking filter

Systems and methods are provided for improving efficiency of decoding performance of a video compression system by selectively applying, based on the size of the coding unit, a prediction method, a palette mode, or palette-based residual refinement method which combines the palette mode and the traditional prediction methods comprising deriving a prediction block of a coding unit, decoding palette-based residual information of the coding unit, deriving a palette-based residual block based on the decoded palette-based residual information, and combining the prediction block and the palette-based residual block to construct a reconstructed block.

Growing community of inventors

Bellevue, WA, United States of America

Jian Lou