公开(公告)号：US20250071299A1

公开(公告)日：2025-02-27

申请号：US18455298

申请日：2023-08-24

Applicant: GOOGLE LLC

Inventor： Yao-Chung Lin ,  Jingning Han ,  Yilin Wang ,  Yeping Su

IPC: H04N19/196 ,  H04N19/105 ,  H04N19/176 ,  H04N19/80

Abstract: Encoding using media compression and processing for machine-learning-based quality metrics includes generating encoded frame data by encoding a current frame from an input video using a neural-network-based video quality model, which includes identifying optimal encoding parameters for encoding a current block, wherein the optimal encoding parameters minimize a rate-distortion optimization cost function, which includes using a gradient value for the current block obtained from a neural-network-based video quality model generated gradient map obtained from the neural-network-based video quality model for the current frame, obtaining a restoration filtered reconstructed frame by restoration filtering a reconstructed frame, obtained by decoding the encoded frame data, using the neural-network-based video quality model generated gradient map obtained for the reconstructed frame.

公开(公告)号：US20240214607A1

公开(公告)日：2024-06-27

申请号：US18594121

申请日：2024-03-04

Applicant: GOOGLE LLC

Inventor： Bohan Li ,  Ching-Han Chiang ,  Jingning Han ,  Yao Yao

IPC: H04N19/597 ,  H04N19/139 ,  H04N19/167 ,  H04N19/176 ,  H04N19/182 ,  H04N19/52 ,  H04N19/57 ,  H04N19/583 ,  H04N19/80

CPC classification number: H04N19/597 ,  H04N19/139 ,  H04N19/167 ,  H04N19/176 ,  H04N19/182 ,  H04N19/52 ,  H04N19/57 ,  H04N19/583 ,  H04N19/80

Abstract: Mapping-aware coding tools for 360 degree videos adapt conventional video coding tools for 360 degree video data using parameters related to a spherical projection of the 360 degree video data. The mapping-aware coding tools perform motion vector mapping techniques, adaptive motion search pattern techniques, adaptive interpolation filter selection techniques, and adaptive block partitioning techniques. Motion vector mapping includes calculating a motion vector for a pixel of a current block by mapping the location of the pixel within a two-dimensional plane (e.g., video frame) onto a sphere and mapping a predicted location of the pixel on the sphere determined based on rotation parameters back onto the plane. Adaptive motion searching, adaptive interpolation filter selection, and adaptive block partitioning operate according to density distortion based on locations along the sphere. These mapping-aware coding tools contemplate changes to video information by the mapping of 360 degree video data into a conventional video format.

公开(公告)号：US11647223B2

公开(公告)日：2023-05-09

申请号：US17132065

申请日：2020-12-23

Applicant: GOOGLE LLC

Inventor： Jingning Han ,  James Bankoski ,  Yaowu Xu

IPC: H04N19/00 ,  H04N19/573 ,  H04N19/70 ,  H04N19/567 ,  H04N19/52

CPC classification number: H04N19/573 ,  H04N19/52 ,  H04N19/567 ,  H04N19/70

Abstract: Dynamic motion vector referencing is used to predict motion within video blocks. A motion trajectory is determined for a current frame including a video block to encode or decode based on a reference motion vector used for encoding or decoding one or more reference frames of the current frame. One or more temporal motion vector candidates are then determined for predicting motion within the video block based on the motion trajectory. A motion vector is selected from a motion vector candidate list including the one or more temporal motion vector candidates and used to generate a prediction block. The prediction block is then used to encode or decode the video block. The motion trajectory is based on an order of video frames indicated by frame offset values encoded to a bitstream. The motion vector candidate list may include one or more spatial motion vector candidates.

公开(公告)号：US11405646B2

公开(公告)日：2022-08-02

申请号：US17106898

申请日：2020-11-30

Applicant: GOOGLE LLC

Inventor： Jingning Han ,  James Zern ,  Linfeng Zhang ,  Ching-Han Chiang ,  Yaowu Xu

IPC: H04N19/60 ,  H04N19/129 ,  H04N19/88 ,  H04N19/184 ,  H04N19/13

Abstract: A processor is configured to maintain, for encoding current values related to the transform coefficients a first line buffer and a second line buffer. The current values are arranged along a current scan-order anti-diagonal line. The first line buffer includes first values of a first scan-order anti-diagonal line. The second line buffer includes second values of a second scan-order anti-diagonal line. The processor is further configured to interleave the first values and the second values in a destination buffer; select, using the destination buffer, a probability distribution for coding a current value of the current values; entropy encode, in a compressed bitstream, the current value using the probability distribution; and replace, for coding values of an immediately subsequent scan-order anti-diagonal line to the current scan-order anti-diagonal line, one of the second line buffer or the first line buffer with the current scan-order anti-diagonal line.

公开(公告)号：US11095890B2

公开(公告)日：2021-08-17

申请号：US16858805

申请日：2020-04-27

Applicant: GOOGLE LLC

Inventor： Cheng Chen ,  Jingning Han

IPC: H04N19/117 ,  H04N19/172 ,  H04N19/426

Abstract: A memory-efficient filtering approach is used to code images and video. A buffer having a fixed size based on a size of processing units to use for filtering a video frame is allocated. For each of the processing units, pre-filtered pixel values are copied from a respective region of the video frame to the buffer based on a writing point for the video frame and an offset applied to the writing point, filtering is performed against the pre-filtered pixel values from the buffer to produce filtered pixel values, and the filtered pixel values are written to the video frame based on the writing point and the offset. The filtering may be performed using a loop restoration tool, such as where the pre-filtered pixel values are output from a constrained directional enhancement filter (CDEF) tool. Alternatively, the filtering may be performed using the CDEF tool or another coding tool.

公开(公告)号：US10750171B2

公开(公告)日：2020-08-18

申请号：US16016768

申请日：2018-06-25

Applicant: GOOGLE LLC

Inventor： Cheng Chen ,  Jingning Han ,  Yaowu Xu

IPC: H04N19/117 ,  H04N19/186 ,  H04N19/46 ,  H04N19/176

Abstract: Systems and methods are disclosed for encoding and decoding video. For example, methods may include: accessing an encoded bitstream; reconstructing an image including multiple color planes based on data from the encoded bitstream; decoding a first filter level from the encoded bitstream, wherein the first filter level specifies one or more thresholds that are used to select a length for a deblocking filter; decoding a second filter level from the encoded bitstream, wherein the second filter level specifies one or more thresholds that are used to select a length for a deblocking filter; after reconstruction of the image, applying a deblocking filter to a first color plane of the image using the first filter level; and, after reconstruction of the image, applying a deblocking filter to a second color plane of the image using the second filter level.

公开(公告)号：US10735767B2

公开(公告)日：2020-08-04

申请号：US16299436

申请日：2019-03-12

Applicant: GOOGLE LLC

Inventor： Yaowu Xu ,  Jingning Han ,  Ching-Han Chiang

IPC: H04N19/129 ,  H04N19/60 ,  H04N19/44 ,  H04N19/18 ,  H04N19/88 ,  G06K9/00 ,  G08B13/196 ,  H04N5/232 ,  H04N19/48 ,  H04N19/645 ,  H04N19/70 ,  H04N19/13 ,  H04N19/91 ,  H04N19/157

Abstract: Encoding a transform block includes de-composing transform coefficients of the transform block into binary level maps arranged in a tier and a residual transform map, the binary level maps formed by breaking down a value of a respective transform coefficient into a series of binary decisions; and encoding, using a context model, a to-be-encoded binary decision that is at a scan location in a scan order, the to-be-encoded binary decision being a value of a binary level map at a level k. The context model is selected using first neighboring binary decisions of the binary level map at a level k that precede the to-be-encoded binary decision; and second neighboring binary decisions of a binary level map at a level (k−1), the second neighboring binary decisions including values that precede and values that follow, in the scan order, a co-located binary decision of the to-be-encoded binary decision.

公开(公告)号：US20200092576A1

公开(公告)日：2020-03-19

申请号：US16131133

申请日：2018-09-14

Applicant: GOOGLE LLC

Inventor： Bohan Li ,  Yaowu Xu ,  Jingning Han

IPC: H04N19/52 ,  H04N19/176 ,  H04N19/577

Abstract: Video coding may include generating, by a processor executing instructions stored on a non-transitory computer-readable medium, an encoded frame by encoding a current frame from an input bitstream, by generating a reference coframe spatiotemporally corresponding to the current frame, wherein the current frame is a frame from a sequence of input frames, wherein each frame from the sequence of input frames has a respective sequential location in the sequence of input frames, and wherein the current frame has a current sequential location in the sequence of input frames, and encoding the current frame using the reference coframe. Video coding may include including the encoded frame in an output bitstream and outputting the output bitstream.

公开(公告)号：US20190342572A1

公开(公告)日：2019-11-07

申请号：US16511389

申请日：2019-07-15

Applicant: GOOGLE LLC

Inventor： Yaowu Xu ,  Jingning Han

IPC: H04N19/52 ,  G06T7/20 ,  G06T9/00 ,  H04N19/159 ,  H04N19/176 ,  H04N19/527 ,  H04N19/44 ,  H04N19/109 ,  H04N19/139

Abstract: Candidate motion vectors are identified from previously coded blocks. For each candidate, a distance between a previously coded block associated with the candidate and a current block and a popularity value based previously coded pixels associated with the candidate are determined. Each candidate is ranked within a list according to a priority value, the priority value based on the distance and the popularity value associated with the candidate. The ranking also includes determining reference blocks that share first and second reference frames for decoding the current block using compound prediction, and assigning a higher priority to candidates that share the same combination of the reference frames within the list than combinations of candidates encoded using respective single reference frames of the first reference frame and the second reference frame. The reference motion vector is selected from the list and is used to decode the current block.

公开(公告)号：US10277905B2

公开(公告)日：2019-04-30

申请号：US14853215

申请日：2015-09-14

Applicant: Google LLC

Inventor： Jingning Han ,  Yaowu Xu ,  Cheng Chen

IPC: H04N7/12 ,  H04N11/02 ,  H04B1/66 ,  H04N19/159 ,  H04N19/176 ,  H04N19/61 ,  H04N19/89 ,  H04N19/625 ,  H04N19/91 ,  H04N19/593 ,  H04N19/124 ,  H04N19/147 ,  H04N19/60 ,  H04N19/112

Abstract: Blocks of a frame of a non-baseband video signal are encoded by selecting between a staircase- or sinusoidal-based transform. For each prediction mode, a prediction block for encoding a current block is generated, a residual block is generated, the residual block is transformed using a sinusoidal-based transform and using a staircase-based transform, and first and second rate-distortion value for encoding the residual block using the sinusoidal-based transform and staircase-based transform are respectively generated. One of the staircase-based transform or the sinusoidal-based transform for encoding the residual block is selected based on a lowest value of the first rate-distortion value and the second rate-distortion value among results for each prediction mode, and the current block is encoded using the selected one of the staircase-based transform or the sinusoidal-based transform. Improved appearance for blocks having sharp edges can be achieved and lossless encoding is possible using the staircase-based transform.