公开(公告)号：US20240323361A1

公开(公告)日：2024-09-26

申请号：US18678031

申请日：2024-05-30

Applicant: GOOGLE LLC

Inventor： Lester Lu ,  Debargha Mukherjee ,  Elliott Karpilovsky

IPC: H04N19/107 ,  H04N19/122 ,  H04N19/176 ,  H04N19/50

CPC classification number: H04N19/107 ,  H04N19/122 ,  H04N19/176 ,  H04N19/50

Abstract: Decoding video data includes, for a block encoded using a prediction mode, determining a transform mode for the block using the prediction mode. The transform mode is a first mode when the prediction mode is an inter-prediction mode and is a second mode when the prediction mode is an intra-prediction mode. The first mode is an available first transform type that is a combination of transforms selected from first fixed transforms and first learned transforms that each comprise a respective transformation matrix generated iteratively using blocks predicted using the inter-prediction mode. The second mode is an available second transform type that is a combination of transforms selected from second fixed transforms, which is a proper subset of the first fixed transforms, and a second learned transform comprising a transformation matrix that is generated iteratively using blocks predicted using the intra-prediction mode. Decoding the block uses the prediction and transform modes.

公开(公告)号：US20240205458A1

公开(公告)日：2024-06-20

申请号：US18542850

申请日：2023-12-18

Applicant: GOOGLE LLC

Inventor： Onur Guleryuz ,  Zeyu Deng ,  Debargha Mukherjee ,  Lester Lu ,  Yue Chen

IPC: H04N19/61 ,  H04N19/105 ,  H04N19/12 ,  H04N19/124 ,  H04N19/13 ,  H04N19/172 ,  H04N19/176 ,  H04N19/42

CPC classification number: H04N19/61 ,  H04N19/105 ,  H04N19/12 ,  H04N19/124 ,  H04N19/13 ,  H04N19/172 ,  H04N19/176 ,  H04N19/42 ,  H04N19/70

Abstract: Transform prediction with parsing independent coding includes generating a reconstructed frame and outputting the reconstructed frame. Generating the reconstructed frame includes entropy decoding transform blocks for the reconstructed frame, entropy decoding decoded transform identifiers for the transform blocks, obtaining transform-specific probability distributions for available transforms, and, for a current transform block from the transform blocks, identifying a current remapped transform identifier from the decoded transform identifiers, identifying a current transform identifier in accordance with the current remapped transform identifier, the transform coefficients from the current transform block, and the transform-specific probability distributions, identifying a current transform in accordance with the current transform identifier; inverse transforming, in accordance with the current transform, the current transform block to obtain a current residual block and obtaining a current reconstructed block using the current residual block. Generating the reconstructed frame includes including the current reconstructed block in the reconstructed frame.

公开(公告)号：US11197004B1

公开(公告)日：2021-12-07

申请号：US16919507

申请日：2020-07-02

Applicant: Google LLC

Inventor： Rohit Singh ,  Debargha Mukherjee ,  Elliott Karpilovsky ,  Lester Lu

IPC: H04N19/184 ,  H04N19/124 ,  H04N19/61

Abstract: Transform modes are derived for inter-predicted blocks using side information available within a bitstream. An inter-predicted encoded video block and side information are identified within a bitstream. Based on the side information, a trained transform is determined for inverse transforming transform coefficients of the inter-predicted encoded video block from amongst multiple trained transforms. The transform coefficients of the inter-predicted encoded video block are inverse transformed according to the trained transform to produce a prediction residual. A video block is reconstructed using the prediction residual and the reference frame. The video block is then output within an output video stream for storage or display. To determine the trained transforms, a learning model uses individual side information types and combinations of the individual side information types processed against a training data set.