Deep Learning-based Video Person Re-identification

In recent years,with the development of the information society and the people's increasing emphasis on public safety,video surveillance equipment has been widely deployed in public places in cities.Person re-identification as a key technology of video surveillance system has become a research hotspot in the field of computer vision.Person re-identification faces problems such as illumination,viewing angle,occlusion,and video person re-identification is the challenge of how to use the timing information of the video sequence for re-identification.The application of deep learning to the technical research of video person re-recognition can effectively process time sequence information and improve the accuracy of re-recognition,which has a certain research value.Aiming at the problem that the current video person re-recognition model has poor effect on the temporal and spatial information extraction of video sequences,a person re-recognition network model based on non-local attention modules and multiple feature fusion is designed,and non-local re-recognition models are embedded in the Res Net-50 basic network.The local attention module extracts global features,and at the same time constructs a multi-layer feature fusion network to obtain the salient features of pedestrians,and finally measures the similarity of pedestrian features and matches and sorts them to obtain the accuracy value.The performance of the proposed model has been significantly improved on each dataset: the values of m AP and Rank-1 on the large dataset MARS are 81.4% and 88.7%,and the values of m AP and Rank-1 on Duke MTMC-Video Re ID reach 93.4% and 95.3%.Moreover,the Rank-1 value on the small dataset PRID2011 is 94.8%.Aiming at the problem that the image generated by the data enhancement method in video person re-identification contains noise interference and cannot extract significant features,a blind denoising and self-supervised compression generation confrontation network is proposed to enhance the data of video person re-identification,using blind denoising generation The confrontation network enhances the original dataset to expand the training dataset,and at the same time denoises the generated pedestrian images.In addition,the self-supervised compression technology is used to compress the generated confrontation network to reduce the amount of calculation.The experimental results demonstrate that the proposed model achieves great performance of Rank-1 reaching 89.1% and 96.7%,m AP value Reached 82.4% and 94.1% on the MARS and Duke MTMC-Video Re ID datasets respectively.