An Efficient Person Search Method Using Spatio-Temporal Features for Surveillance Videos

Existing person search methods mainly focus on searching for the target using database images. However, this is different from real-world surveillance videos which involve a temporal relationship between video frames. To solve problem, we propose an efficient method that employs spatio-temporal features in videos. This not only considers spatial of persons each frame, but also utilizes same adjacent For purpose, are extracted by combining Yolo network with Resnet-50 model, and processed gated recurrent unit. The generated following average pooling layer used to represent ensure efficiency, locality sensitive hashing organize massive calculate similarity. A constructed evaluate proposed method, experimental results demonstrate our improves accuracy while ensuring efficiency.