Multi-Target Multi-Camera Tracking Based On Mutual Information-Temporal Weight Aggregation Person Re-Identification
Jiayue Li, Yan Piao · 2022 IEEE 5th International Conference on Electronic Information and Communication Technology (ICEICT) · 2022
In order to solve the problem of multi-target multi-camera tracking, the method of detection and tracking is used to complete single-camera multi-target tracking, and a person re-identification model using mutual information time weight aggregation is proposed to extract appearance features, and determine whether the camera is in the camera. A new method for cross-camera multi-target tracking is finally performed. We use the deep learning framework Pytorch to build network models and tune model-related parameters and compare experiments with existing models on the dataset. The results show that the model has better tracking performance.