Person re-identification is a cross-view pedestrian tracking and retrieval technology, which is of great significance in the field of security monitoring. Due to the different conditions of the shooting scene, there will be a series of problems such as low resolution, perspective occlusion, pose changes, and lighting, which bring many challenges to the application of person re-identification technology. In order to solve the problem of low model recognition accuracy due to the loss of image block information and insufficient expression of pedestrian local features in person re-identification, this paper proposes a person re-identification method based on improved Transformer and CNN. Using the backbone network combined with ResNet and Transformer enhances the ability of pedestrian feature extraction. Through a large number of experiments on the mainstream data sets Market1501 and DukeMTMC-reID, the experimental results show that the person re-identification algorithm based on the improved Transformer and CNN can effectively improve the accuracy of person re-identification
Person Re-Identification Based on Improved Transformer and CNN
2023-10-11
2123201 byte
Conference paper
Electronic Resource
English
Single-person tool-free moving device for heavy transformer
European Patent Office | 2021
|Person Identification System for UAV
Springer Verlag | 2021
|