公开(公告)号：US12169783B2

公开(公告)日：2024-12-17

申请号：US17586136

申请日：2022-01-27

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Qixiang Ye ,  Tianliang Zhang ,  Jianzhuang Liu ,  Xiaopeng Zhang ,  Qi Tian ,  Lihui Jiang

IPC: G06N3/084 ,  G06F18/25 ,  G06N3/045 ,  G06V10/20 ,  G06V10/77 ,  G06V10/80 ,  G06V10/82 ,  G06V20/40 ,  G06V20/52 ,  G06V20/58 ,  G06V40/10

Abstract: This application relates to the field of artificial intelligence and the field of computer vision. The method includes performing feature extraction on an image to obtain a basic feature map of the image, and determining a proposal of a region possibly including a pedestrian in the image. The basic feature map of the image is then processed to obtain an object visibility map in which a response to a pedestrian visible part is greater than a response to a pedestrian blocked part and a background part. The method further performs weighted summation processing on the object visibility map and the basic feature map to obtain an enhanced feature map of the image, and determines, based on the proposal of the image and the enhanced feature map of the image, a bounding box including a pedestrian in the image and a confidence level of the bounding box including the pedestrian.

公开(公告)号：US12026938B2

公开(公告)日：2024-07-02

申请号：US17573220

申请日：2022-01-11

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Yuhui Xu ,  Lingxi Xie ,  Xiaopeng Zhang ,  Xin Chen ,  Guojun Qi ,  Qi Tian

IPC: G06V10/82 ,  G06N3/082 ,  G06N3/10 ,  G06V10/764

CPC classification number: G06V10/82 ,  G06N3/082 ,  G06N3/10 ,  G06V10/764

Abstract: Example neural architecture search methods and image processing methods and apparatuses in the field of computer vision in the field of artificial intelligence are provided. The example neural architecture search method includes determining search space and a plurality of construction units, superimposing the plurality of construction units to obtain a search network, adjusting, in the search space, network architectures of the construction units in the search network, to obtain optimized construction units, and establishing a target neural network based on the optimized construction units. In each construction unit, some channels of an output feature map of each node are processed by using a to-be-selected operation to obtain a processed feature map, and the processed feature map and a remaining feature map are stitched and then input to a next node.

公开(公告)号：US11132780B2

公开(公告)日：2021-09-28

申请号：US16791806

申请日：2020-02-14

Applicant: Huawei Technologies Co., Ltd.

Inventor： Lingxi Xie ,  Kaiwen Duan ,  Qi Tian

IPC: G06K9/00 ,  G06T7/00 ,  G06T7/70 ,  G06T7/80 ,  G06K9/46 ,  G01S17/06 ,  G06N3/08 ,  G06N20/10 ,  G01S17/89

Abstract: A target detection method comprises performing corner point feature extraction processing on an input image to obtain a pair of target corner points, generating a target frame based on the pair of target corner points, and outputting a target detection result of a target object when determining that a calibration area in the target frame comprises a target feature point.

公开(公告)号：US20210256680A1

公开(公告)日：2021-08-19

申请号：US16791806

申请日：2020-02-14

Applicant: Huawei Technologies Co., Ltd.

Inventor： Lingxi Xie ,  Kaiwen Duan ,  Qi Tian

IPC: G06T7/00 ,  G06T7/70 ,  G06T7/80 ,  G06K9/46 ,  G01S17/89 ,  G01S17/06 ,  G06N3/08 ,  G06N20/10

Abstract: A target detection method comprises performing corner point feature extraction processing on an input image to obtain a pair of target corner points, generating a target frame based on the pair of target corner points, and outputting a target detection result of a target object when determining that a calibration area in the target frame comprises a target feature point.