公开(公告)号：US20230386241A1

公开(公告)日：2023-11-30

申请号：US18088800

申请日：2022-12-27

Applicant: UBTECH ROBOTICS CORP LTD

Inventor： KAN WANG ,  Jianxin Pang

IPC: G06V40/10 ,  G06V10/74

CPC classification number: G06V40/10 ,  G06V10/761

Abstract: A person re-identification method, a storage medium, and a terminal device are provided. In the method, a loss function used during model training is a preset distribution-based triplet loss function constraining a difference between a mean of a negative sample feature distance and a mean of a positive sample feature distance to be larger than a preset difference threshold; where the positive sample feature distance is a distance between a feature of a reference image, and a feature of a positive sample image, and the negative sample feature distance is a distance between the feature of the reference image and a feature of a negative sample image. In this manner, it can constrain the mean of the positive sample feature distance and that of the negative sample feature distance, thereby improving the accuracy of person re-identification results.

公开(公告)号：US20240193930A1

公开(公告)日：2024-06-13

申请号：US18536289

申请日：2023-12-12

Applicant: UBTECH ROBOTICS CORP LTD

Inventor： KAN WANG ,  Jiansin Pang

IPC: G06V10/82 ,  G06V10/77

CPC classification number: G06V10/82 ,  G06V10/7715

Abstract: A target identification method includes: obtaining an image containing a target to be identified; performing feature extraction on the image to obtain image features in the image; and inputting the image features into a target identification network model to obtain an identification result that determines a class to which the target to be identified belongs. The target recognition network model includes a loss function that is to constrain a first distance corresponding to each triplet in a number of triplets and a second distance corresponding to each triplet in the triplets. The first distance represents a distance between anchor image features and positive sample image features in each triplet in the triplets, and the second distance represents a distance between the anchor image features and a class center mean of a number of classes in each triplet in the triplets.

公开(公告)号：US20240193929A1

公开(公告)日：2024-06-13

申请号：US18536287

申请日：2023-12-12

Applicant: UBTECH ROBOTICS CORP LTD

Inventor： KAN WANG ,  Jianxin Pang ,  Huan Tan

IPC: G06V10/82 ,  G06V10/77

CPC classification number: G06V10/82 ,  G06V10/7715 ,  G06V2201/07

Abstract: A target identification method includes: obtaining an image containing a target to be identified; performing feature extraction on the image to obtain image features in the image; and inputting the image features into a target identification network model to obtain an identification result that determines a class to which the target to be identified belongs. The target identification network model includes a loss function that is based on intra-class constraints and inter-class constraints. The intra-class constraints are to constrain an intra-class distance between sample image features of a sample target and a class center of a class to which the sample target belongs, and the inter-class constraints are to constrain inter-class distances between class centers of different classes, and/or inter-class angles between the class centers of different classes.