公开(公告)号：US20240169017A1

公开(公告)日：2024-05-23

申请号：US18425726

申请日：2024-01-29

Applicant: Imagination Technologies Limited

Inventor： Cagatay Dikici ,  Clifford Gibson ,  James Imber

IPC: G06F17/15 ,  G06N3/063

CPC classification number: G06F17/153 ,  G06N3/063

Abstract: Methods and systems for performing a convolution transpose operation between an input tensor having a plurality of input elements and a filter comprising a plurality of filter weights. The method includes: dividing the filter into a plurality of sub-filters; performing, using hardware logic, a convolution operation between the input tensor and each of the plurality of sub-filters to generate a plurality of sub-output tensors, each sub-output tensor comprising a plurality of output elements; and interleaving, using hardware logic, the output elements of the plurality of sub-output tensors to form a final output tensor for the convolution transpose.

公开(公告)号：US20240161253A1

公开(公告)日：2024-05-16

申请号：US18478050

申请日：2023-09-29

Applicant: Imagination Technologies Limited

Inventor： James Imber ,  Joseph Heyward ,  Kristof Beets ,  John Viljoen

IPC: G06T5/73 ,  G06T3/4007 ,  G06T3/4053

CPC classification number: G06T5/73 ,  G06T3/4007 ,  G06T3/4053 ,  G06T2207/20212

Abstract: Methods and processing modules apply adaptive sharpening, for a block of input pixels for which upsampling is performed, to determine a block of output pixels. A block of upsampled pixels is obtained based on the block of input pixels. One or more range kernels is determined based on a plurality of upsampled pixels of the block of upsampled pixels. Each of the one or more range kernels is combined with a sharpening kernel to determine one or more bilateral sharpening kernels. The one or more bilateral sharpening kernels are used to determine the output pixels of the block of output pixels.

公开(公告)号：US20240135505A1

公开(公告)日：2024-04-25

申请号：US18373710

申请日：2023-09-27

Applicant: Imagination Technologies Limited

Inventor： James Imber ,  Joseph Heyward ,  Kristof Beets

IPC: G06T5/73 ,  G06T3/4053

CPC classification number: G06T5/73 ,  G06T3/4053

Abstract: Methods and processing modules apply adaptive sharpening, for a block of input pixels, to determine a block of output pixels. A block of sharp pixels is obtained based on the block of input pixels, the block of sharp pixels being for representing a sharp version of the block of output pixels. One or more indications of contrast for the block of input pixels is determined. Each of the output pixels of the block of output pixels is determined by performing a respective weighted sum of: (i) a corresponding input pixel in the block of input pixels and (ii) a corresponding sharp pixel in the block of sharp pixels. The weights of the weighted sums are based on the determined one or more indications of contrast for the block of input pixels.

公开(公告)号：US11915397B2

公开(公告)日：2024-02-27

申请号：US17956925

申请日：2022-09-30

Applicant: Imagination Technologies Limited

Inventor： Szabolcs Csefalvay ,  James Imber ,  David Walton ,  Insu Yu

IPC: G06T5/00 ,  G06T5/20 ,  G06T5/50 ,  G06T15/06 ,  G06T15/50 ,  G06T3/40

CPC classification number: G06T5/002 ,  G06T3/40 ,  G06T5/20 ,  G06T5/50 ,  G06T15/06 ,  G06T15/506 ,  G06T2207/20016 ,  G06T2207/20212 ,  G06T2207/20221

Abstract: A method of rendering an image of a 3-D scene includes rendering a noisy image; and obtaining one or more guide channels. For each of a plurality of local neighbourhoods, the method comprises: calculating the parameters of a model that approximates the noisy image as a function of the one or more guide channels, and applying the calculated parameters to produce a denoised image. Tiling is used when calculating the parameters of the model.

公开(公告)号：US20230306248A1

公开(公告)日：2023-09-28

申请号：US18132929

申请日：2023-04-10

Applicant: Imagination Technologies Limited

Inventor： Clifford Gibson ,  James Imber

IPC: G06N3/063 ,  G06F7/544 ,  G06F7/00 ,  G06N3/045 ,  G06N3/04 ,  G06N3/08

CPC classification number: G06N3/063 ,  G06F7/00 ,  G06F7/5443 ,  G06N3/04 ,  G06N3/045 ,  G06N3/08

Abstract: A method in a hardware implementation of a Convolutional Neural Network (CNN), includes receiving a first subset of data having at least a portion of weight data and at least a portion of input data for a CNN layer and performing, using at least one convolution engine, a convolution of the first subset of data to generate a first partial result; receiving a second subset of data comprising at least a portion of weight data and at least a portion of input data for the CNN layer and performing, using the at least one convolution engine, a convolution of the second subset of data to generate a second partial result; and combining the first partial result and the second partial result to generate at least a portion of convolved data for a layer of the CNN.

公开(公告)号：US20230117042A1

公开(公告)日：2023-04-20

申请号：US17968025

申请日：2022-10-18

Applicant: Imagination Technologies Limited

Inventor： Sandra La Mantia ,  Cagatay Dikici ,  James Imber ,  Timothy Atherton

IPC: G06F17/14 ,  G06F17/15

Abstract: A mechanism for performing a discrete Fourier-related transform using a hardware accelerator that comprises fixed-function circuitry including convolution hardware configured to perform one or more convolution operations. A matrix multiplication operation used in the discrete Fourier-related transform is performed by the convolution hardware using a convolution operation. A convolution kernel for the convolution operation is derived from a weight matrix representing a multiplicand or multiplier of the matrix multiplication operation.

公开(公告)号：US11625581B2

公开(公告)日：2023-04-11

申请号：US15585571

申请日：2017-05-03

Applicant: Imagination Technologies Limited

Inventor： Clifford Gibson ,  James Imber

IPC: G06N3/063 ,  G06N3/04 ,  G06F7/544 ,  G06F7/00 ,  G06N3/08

Abstract: A method in a hardware implementation of a Convolutional Neural Network (CNN), includes receiving a first subset of data having at least a portion of weight data and at least a portion of input data for a CNN layer and performing, using at least one convolution engine, a convolution of the first subset of data to generate a first partial result; receiving a second subset of data comprising at least a portion of weight data and at least a portion of input data for the CNN layer and performing, using the at least one convolution engine, a convolution of the second subset of data to generate a second partial result; and combining the first partial result and the second partial result to generate at least a portion of convolved data for a layer of the CNN.

公开(公告)号：US20220327366A1

公开(公告)日：2022-10-13

申请号：US17846803

申请日：2022-06-22

Applicant: Imagination Technologies Limited

Inventor： James Imber

IPC: G06N3/063 ,  G06N3/08 ,  G06F7/483 ,  G06F17/11

Abstract: Methods for determining a fixed point format for one or more layers of a DNN based on the portion of the output error of the DNN attributed to the fixed point formats of the different layers. Specifically, in the methods described herein the output error of a DNN attributable to the quantisation of the weights or input data values of each layer is determined using a Taylor approximation and the fixed point number format of one or more layers is adjusted based on the attribution. For example, where the fixed point number formats used by a DNN comprises an exponent and a mantissa bit length, the mantissa bit length of the layer allocated the lowest portion of the output error may be reduced, or the mantissa bit length of the layer allocated the highest portion of the output error may be increased. Such a method may be iteratively repeated to determine an optimum set of fixed point number formats for the layers of a DNN.

公开(公告)号：US20220253716A1

公开(公告)日：2022-08-11

申请号：US17568325

申请日：2022-01-04

Applicant: Imagination Technologies Limited

Inventor： Biswarup Choudhury ,  Aria Ahmadi ,  James Imber ,  Cagatay Dikici ,  Timothy Atherton

IPC: G06N3/10 ,  G06N3/04 ,  G06F17/16 ,  G06V10/82

Abstract: A method and data processing system implement a neural network containing at least one matrix multiplication operation. The matrix multiplication operation is mapped to a graph of neural network operations including at least one transformation and at least one convolution. The at least one convolution is implemented in fixed-function hardware of a neural network accelerator.

公开(公告)号：US20220027717A1

公开(公告)日：2022-01-27

申请号：US17498618

申请日：2021-10-11

Applicant: Imagination Technologies Limited

Inventor： Clifford Gibson ,  James Imber

IPC: G06N3/063 ,  G06N3/04 ,  G06N3/08

Abstract: A method of configuring a hardware implementation of a Convolutional Neural Network (CNN), the method comprising: determining, for each of a plurality of layers of the CNN, a first number format for representing weight values in the layer based upon a distribution of weight values for the layer, the first number format comprising a first integer of a first predetermined bit-length and a first exponent value that is fixed for the layer; determining, for each of a plurality of layers of the CNN, a second number format for representing data values in the layer based upon a distribution of expected data values for the layer, the second number format comprising a second integer of a second predetermined bit-length and a second exponent value that is fixed for the layer; and storing the determined number formats for use in configuring the hardware implementation of a CNN.