公开(公告)号：US20240020519A1

公开(公告)日：2024-01-18

申请号：US18351417

申请日：2023-07-12

Applicant: CANON KABUSHIKI KAISHA

Inventor： Wei Tao ,  Tsewei Chen ,  Deyu Wang ,  Lingxiao Yin ,  Dongyue Zhao

IPC: G06N3/0495 ,  G06N3/084

CPC classification number: G06N3/0495 ,  G06N3/084

Abstract: The present disclosure provides training and application methods and apparatuses for a neural network model, and a storage medium. The training method includes: quantizing, in a forward transfer process, a network parameter represented by a continuous real value, and calculating a quantization error; determining, in a backward transfer process, a gradient of a weight in the neural network model; correcting the gradient of the weight based on the calculated quantization error, wherein the correcting includes correcting a magnitude of the gradient and correcting a direction of the gradient; and updating the neural network model according to the corrected gradient.

公开(公告)号：US20230154174A1

公开(公告)日：2023-05-18

申请号：US18055603

申请日：2022-11-15

Applicant: CANON KABUSHIKI KAISHA

Inventor： Tsewei Chen ,  Masami Kato

IPC: G06V10/82 ,  G06F7/544

CPC classification number: G06V10/82 ,  G06F7/5443

Abstract: An apparatus for performing filter processing on a data array in a processing target block of a predetermined size is provided. A data memory holds the data array in the processing target block. A coefficient memory holds weight coefficients of a filter used for the filter processing. A controller determines, in a determination, whether data in a reference region in the processing target block, set in correspondence with the processing target block, are zero values. A processor generates a convolution operation result of the weight coefficients and data at a plurality of positions in the processing target block. The controller controls, based on a result of the determination, whether to perform at least some of multiply-accumulate operations of the data and the weight coefficients when the processor generates the convolution operation result.

公开(公告)号：US20220392207A1

公开(公告)日：2022-12-08

申请号：US17825962

申请日：2022-05-26

Applicant: CANON KABUSHIKI KAISHA

Inventor： Masami Kato ,  Shiori Wakino ,  Tsewei Chen ,  Kinya Osa ,  Motoki Yoshinaga

IPC: G06V10/82 ,  G06V10/75 ,  G06V10/77

Abstract: An information processing apparatus operable to perform computation processing in a neural network comprises a coefficient storage unit configured to store filter coefficients of the neural network, a feature storage unit configured to store feature data, a storage control unit configured to store in the coefficient storage unit a part of previously obtained feature data as template feature data, a convolution operation unit configured to compute new feature data by a convolution operation between feature data stored in the feature storage unit and filter coefficients stored in the coefficient storage unit, and compute, by a convolution operation between feature data stored in the feature storage unit and the template feature data stored in the coefficient storage unit, correlation data between the feature data stored in the feature storage unit and the template feature data.

公开(公告)号：US20220012856A1

公开(公告)日：2022-01-13

申请号：US17353959

申请日：2021-06-22

Applicant: CANON KABUSHIKI KAISHA

Inventor： Tsewei Chen ,  Masami Kato ,  Shiori Wakino

IPC: G06T5/20 ,  G06K9/66 ,  G06T5/10 ,  G06T5/00 ,  G06T3/60

Abstract: There is provided with a processing apparatus. A data holder holds at least some of data of a plurality of channels in a target layer among a plurality of layers. Each of a plurality of processors performs, in parallel, a product-sum operation using the data of one channel of the target layer and a coefficient corresponding to the target layer. A selector selects whether to perform first processing or second processing on the basis of information specifying processing in the target layer. The first processing includes inputting the data of one channel of the target layer into one of the plurality of processors. The second processing includes inputting the data of one channel of the target layer to the plurality of processors in parallel.

公开(公告)号：US20190102887A1

公开(公告)日：2019-04-04

申请号：US16140279

申请日：2018-09-24

Applicant: CANON KABUSHIKI KAISHA

Inventor： Qin Yang ,  Tsewei Chen

IPC: G06T7/194 ,  G06K9/20 ,  G06T7/11

Abstract: An image processing apparatus including a unit configured to acquire a current image from an inputted video and a background model which comprises a background image and foreground/background classification information of visual elements; a unit configured to determine first similarity measures between visual elements in the current image and the visual elements in the background model; and a unit configured to classify the visual elements in the current image as the foreground or the background according to the current image, the background image in the background model and the first similarity measures. Wherein, the visual elements in the background model are the visual elements whose classification information is the background and which neighbour to corresponding portions of the visual elements in the current image. Accordingly, the accuracy of the foreground detection could be improved.

公开(公告)号：US09824301B2

公开(公告)日：2017-11-21

申请号：US15139603

申请日：2016-04-27

Applicant: CANON KABUSHIKI KAISHA

Inventor： Tsewei Chen

IPC: G06K9/62 ,  G06K9/54 ,  G06K9/00

CPC classification number: G06K9/6282 ,  G06K9/00228 ,  G06K9/6257 ,  G06K9/6286

Abstract: In an information processing apparatus that includes sequences of weak classifiers which are logically cascade-connected in each sequence and the sequences respectively correspond to categories of an object and in which the weak classifiers are grouped into at least a first group and a second group in the order of connection, classification processing by weak classifiers belonging to the first group of respective categories is performed by pipeline processing. Based on the processing results of the weak classifiers belonging to the first group of the respective categories, categories in which classification processing by weak classifiers belonging to the second group is to be performed are decided out of the categories. The classification processing by the weak classifiers respectively corresponding to the decided categories and belonging to the second group is performed by pipeline processing.

公开(公告)号：US12147901B2

公开(公告)日：2024-11-19

申请号：US16721624

申请日：2019-12-19

Applicant: CANON KABUSHIKI KAISHA

Inventor： Hongxing Gao ,  Wei Tao ,  Tsewei Chen ,  Dongchao Wen ,  Junjie Liu

IPC: G06N3/04 ,  G06N3/084

Abstract: The present disclosure provides a training and application method of a multi-layer neural network model, apparatus and a storage medium. In a forward propagation of the multi-layer neural network model, the number of input feature maps is expanded and a data computation is performed by using the expanded input feature maps.

公开(公告)号：US11900577B2

公开(公告)日：2024-02-13

申请号：US17353959

申请日：2021-06-22

Applicant: CANON KABUSHIKI KAISHA

Inventor： Tsewei Chen ,  Masami Kato ,  Shiori Wakino

IPC: G06T5/20 ,  G06V30/194 ,  G06T3/60 ,  G06T5/00 ,  G06T5/10

CPC classification number: G06T5/20 ,  G06T3/606 ,  G06T5/009 ,  G06T5/10 ,  G06V30/194 ,  G06T2207/20081 ,  G06T2207/20084

Abstract: There is provided with a processing apparatus. A data holder holds at least some of data of a plurality of channels in a target layer among a plurality of layers. Each of a plurality of processors performs, in parallel, a product-sum operation using the data of one channel of the target layer and a coefficient corresponding to the target layer. A selector selects whether to perform first processing or second processing on the basis of information specifying processing in the target layer. The first processing includes inputting the data of one channel of the target layer into one of the plurality of processors. The second processing includes inputting the data of one channel of the target layer to the plurality of processors in parallel.

公开(公告)号：US11853864B2

公开(公告)日：2023-12-26

申请号：US16799199

申请日：2020-02-24

Applicant: CANON KABUSHIKI KAISHA

Inventor： Motoki Yoshinaga ,  Tsewei Chen ,  Masami Kato

IPC: G06N3/063 ,  G06N3/08 ,  G06N3/04 ,  G06F18/213 ,  G06V10/77 ,  G06V10/82

CPC classification number: G06N3/063 ,  G06F18/213 ,  G06N3/04 ,  G06N3/08 ,  G06V10/7715 ,  G06V10/82

Abstract: A data processing apparatus for executing data processing using a neural network including a plurality of hierarchal levels includes an extraction unit configured to extract intermediate feature data from input feature data, a calculation unit configured to calculate output feature data by reducing the number of channels of the intermediate feature data, a storage unit configured to store the output feature data calculated by the calculation unit and provide the input feature data to the extraction unit, and a control unit configured to control the number of channels of the intermediate feature data to be extracted by the extraction unit and the number of channels of the output feature data to be calculated by the calculation unit.

公开(公告)号：US11847569B2

公开(公告)日：2023-12-19

申请号：US16721606

申请日：2019-12-19

Applicant: CANON KABUSHIKI KAISHA

Inventor： Wei Tao ,  Hongxing Gao ,  Tsewei Chen ,  Dongchao Wen ,  Junjie Liu

IPC: G06N20/00 ,  G06N3/084 ,  G06N3/04

CPC classification number: G06N3/084 ,  G06N3/04

Abstract: The present disclosure provides a training and application method of a multi-layer neural network model, apparatus and storage medium. A number of channels of a filter in at least one convolutional layer in the multi-layer neural network model is expanded, and a convolution computation is performed by using the filter after expanding the number of channels, so that the performance of the network model does not degrade while simplifying the network model.