公开(公告)号：US20250008128A1

公开(公告)日：2025-01-02

申请号：US18884321

申请日：2024-09-13

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Timofey Mikhailovich Solovyev ,  Esin Koyuncu ,  Alexander Alexandrovich Karabutov ,  Maxim Borisovitch Sychev ,  Mikhail Vyacheslavovich Sosulnikov ,  Sergey Yurievich Ikonin ,  Elena Alexandrovna Alshina

IPC: H04N19/189 ,  H04N19/13 ,  H04N19/42

Abstract: A neural network including at least one neural network layer and an activation function connected to an output of the at least one neural network layer. The activation function is implemented as an approximation function of a mathematically defined real valued non-linear activation function, wherein the approximation function allows for integer-only processing of fixed-point representations of input values of the approximation function.

公开(公告)号：US20250005331A1

公开(公告)日：2025-01-02

申请号：US18885411

申请日：2024-09-13

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Timofey Mikhailovich Solovyev ,  Sergey Yurievich Ikonin ,  Elena Alexandrovna Alshina ,  Johannes Sauer ,  Esin Koyuncu ,  Maxim Borisovitch Sychev ,  Alexander Alexandrovich Karabutov ,  Mikhail Vyacheslavovich Sosulnikov ,  Kirill Igorevich SOLODSKIKH ,  Vladimir Mikhailovich Kryzhanovskiy ,  Alexander Nikolaevich Filippov

IPC: G06N3/0455 ,  G06T9/00

Abstract: The present disclosure relates to a method of operating a neural network with clipped input data. The method includes defining lower and upper threshold values for integer numbers in data entities of input data for at least one neural network layer. If a value of an integer number in a data entity of the input data is smaller than the defined lower threshold value, the method includes clipping the value of the integer number comprised in the data entity of the input data to the defined lower threshold value. If a value of an integer number in a data entity of the input data is larger than the defined upper threshold value, the method includes clipping the value of the integer number comprised in the data entity of the input data to the defined upper threshold value. Integer overflow of an accumulator register is thereby avoided.

公开(公告)号：US20250005330A1

公开(公告)日：2025-01-02

申请号：US18883907

申请日：2024-09-12

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Timofey Mikhailovich Solovyev ,  Sergey Yurievich Ikonin ,  Elena Alexandrovna Alshina ,  Esin Koyuncu ,  Maxim Borisovitch Sychev ,  Alexander Alexandrovich Karabutov ,  Mikhail Vyacheslavovich Sosulnikov

IPC: G06N3/0455

Abstract: A method of operating a neural network based on conditioned weights includes defining integer lower and upper threshold values for values of integer numbers comprised in data entities of input data for the neural network layer. If a value of an integer numbers comprised in a data entity of the input data is smaller than the lower threshold value, the value of the integer number comprised in the data entity of the input data is clipped to the lower threshold value, or if a value of an integer number comprised in a data entity of the input data is larger than the upper threshold value, the value of the integer number comprised in the data entity of the input data is clipped to the upper threshold value. Integer valued weights are determined based on the lower threshold value, the upper threshold value, and a pre-defined accumulator register size, such that integer overflow of the accumulator register can be avoided.

公开(公告)号：US12160578B2

公开(公告)日：2024-12-03

申请号：US17509324

申请日：2021-10-25

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Sergey Yurievich Ikonin ,  Roman Igorevich Chernyak ,  Timofey Mikhailovich Solovyev ,  Alexander Alexandrovich Karabutov ,  Jianle Chen ,  Elena Alexandrovna Alshina

IPC: H04N19/17 ,  H04N19/124 ,  H04N19/126 ,  H04N19/132 ,  H04N19/172 ,  H04N19/174 ,  H04N19/184 ,  H04N19/186 ,  H04N19/44

Abstract: A method of obtaining a chrominance quantization parameter (QP) for chrominance components based on a luminance QP for luminance component, wherein the method is performed by a decoder, comprising parsing a received bitstream to obtain the luminance QP and information of a mapping function (ƒ) which associates a QP index (QPi) to the chrominance QP (QPc); obtaining the QPi based at least in a part on the luminance QP; obtaining the mapping function based on the obtained information; and obtaining a QPc based on the obtained mapping function and the obtained QPi.

公开(公告)号：US12088820B2

公开(公告)日：2024-09-10

申请号：US17896012

申请日：2022-08-25

Applicant: Huawei Technologies Co., Ltd.

Inventor： Semih Esenlik ,  Biao Wang ,  Anand Meher Kotra ,  Elena Alexandrovna Alshina

IPC: H04N19/593 ,  H04N19/14 ,  H04N19/174 ,  H04N19/423

CPC classification number: H04N19/14 ,  H04N19/174 ,  H04N19/423

Abstract: A method of coding performed by a decoding device and a decoder are disclosed. The method comprises: obtaining a bitstream for a current picture; obtaining a quantity of tiles in a tile row of the current picture; obtaining a value of an address of a current slice, the current slice is comprised in the current picture; obtaining a value of a horizontal coordinate according to the value of the address of the current slice; parsing an indication value for a slice width from the bitsteam for the current picture, when a difference value between the quantity of tiles in the tile row of the current picture and the value of the horizontal coordinate is not equal to a first threshold.

公开(公告)号：US12063388B2

公开(公告)日：2024-08-13

申请号：US17850123

申请日：2022-06-27

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Biao Wang ,  Semih Esenlik ,  Anand Meher Kotra ,  Han Gao ,  Elena Alexandrovna Alshina

IPC: H04N19/593 ,  H04N19/105 ,  H04N19/132 ,  H04N19/172 ,  H04N19/176

CPC classification number: H04N19/593 ,  H04N19/105 ,  H04N19/132 ,  H04N19/172 ,  H04N19/176

Abstract: A method of coding implemented is provided. The method includes the following operations: obtained the height and width of a prediction block without applying clipping operation; calculating a value of a vertical component of an intra prediction sample based on the height and width of the prediction block; calculating a value of a horizontal component of the intra prediction sample based on the height and width of the prediction block; and generating the intra prediction sample based on the value of the vertical component and the value of the horizon component.

公开(公告)号：US12047609B2

公开(公告)日：2024-07-23

申请号：US17707202

申请日：2022-03-29

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Timofey Mikhailovich Solovyev ,  Maxim Borisovitch Sychev ,  Huanbang Chen ,  Alexander Alexandrovich Karabutov ,  Roman Igorevich Chernyak ,  Sergey Yurievich Ikonin ,  Haitao Yang ,  Elena Alexandrovna Alshina

IPC: H04N19/82 ,  H04N19/105 ,  H04N19/109 ,  H04N19/117 ,  H04N19/132 ,  H04N19/139 ,  H04N19/176 ,  H04N19/182 ,  H04N19/196 ,  H04N19/52 ,  H04N19/80

CPC classification number: H04N19/80 ,  H04N19/105 ,  H04N19/109 ,  H04N19/117 ,  H04N19/132 ,  H04N19/139 ,  H04N19/176 ,  H04N19/182 ,  H04N19/197 ,  H04N19/52

Abstract: An apparatus for inter prediction of a sample value of a current full-integer pixel of a plurality of pixels associated with a current block of a current frame of a video signal is disclosed. The apparatus is configured to: determine a motion vector of the current full-integer pixel based on a motion compensation model; determine for the current full-integer pixel a corresponding sub-integer pixel in a reference frame; determine positions of a set of corresponding filter support pixels in the reference frame; obtain a respective sample value of the corresponding sub-integer pixel of the current full-integer pixel and the corresponding filter support pixels in the reference frame by applying a bilinear interpolation; and determine the inter predicted sample value of the current full-integer pixel.

公开(公告)号：US12041230B2

公开(公告)日：2024-07-16

申请号：US17712980

申请日：2022-04-04

Applicant: Huawei Technologies Co., Ltd.

Inventor： Elena Alexandrovna Alshina ,  Semih Esenlik ,  Anand Meher Kotra ,  Biao Wang ,  Han Gao

IPC: H04N19/117 ,  H04N19/132 ,  H04N19/184 ,  H04N19/30 ,  H04N19/80

CPC classification number: H04N19/117 ,  H04N19/132 ,  H04N19/184 ,  H04N19/30 ,  H04N19/80

Abstract: A method operates a two-dimensional (2D) separable interpolation filter for coding a video. The 2D separable interpolation filter includes a first second one-dimensional (1D) interpolation filter and a temporal buffer. The method includes: obtaining filter coefficients for the first 1D interpolation filter; comparing a sum of positive filter coefficients of the obtained filter coefficients with a threshold; and upon determining that the sum of positive filter coefficients is larger than the threshold, amending one or more of the filter coefficients to obtain amended filter coefficients, the filter coefficients being amended such that a sum of positive filter coefficients is not larger than the threshold; applying the amended filter coefficients to samples of the video to obtain a value of a fractional sample position of the video; re-scaling the value of the fractional sample position; and storing the re-scaled value of the fractional sample position in the temporal buffer.

公开(公告)号：US12010291B2

公开(公告)日：2024-06-11

申请号：US17557930

申请日：2021-12-21

Applicant: Huawei Technologies Co., Ltd.

Inventor： Semih Esenlik ,  Max Blaeser ,  Zhijie Zhao ,  Han Gao ,  Biao Wang ,  Anand Meher Kotra ,  Elena Alexandrovna Alshina

IPC: H04N19/103 ,  H04N19/119 ,  H04N19/132 ,  H04N19/176 ,  H04N19/46 ,  H04N19/70

CPC classification number: H04N19/103 ,  H04N19/119 ,  H04N19/132 ,  H04N19/176 ,  H04N19/46 ,  H04N19/70

Abstract: A method for coding implemented by a decoding device or an encoding device includes: obtaining a first prediction mode for a first sub-block, obtaining a second prediction mode for a second sub-block; obtaining a first prediction value for a sample in the current block according to the first prediction mode, obtaining a second prediction value for the sample in the current block according to the second prediction mode; obtaining a value of an indicator, the value of the indicator indicating a blending operation which is performed for the sample in the current block; and obtaining a combined prediction value for the sample in the current block, according to the value of the indictor, the first prediction value, and the second prediction value. The blending operation adaptively performs around a separation line dividing the current block into at least the first sub-block and the second sub-block.

公开(公告)号：US20240078414A1

公开(公告)日：2024-03-07

申请号：US18507949

申请日：2023-11-13

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Ahmet Burakhan Koyuncu ,  Atanas Boev ,  Elena Alexandrovna Alshina

IPC: G06N3/0464 ,  G06N7/01

CPC classification number: G06N3/0464 ,  G06N7/01

Abstract: Methods and apparatuses are provided for entropy encoding and decoding of a latent tensor, which includes separating the latent tensor into patches and obtaining a probability model for the entropy encoding of a current element of the latent tensor by processing a set of elements from different patches by one or more layers of a neural network. The processing of the set of elements by applying a convolution kernel enables sharing of information between the separated patches.