| abstract |
The invention discloses a pedestrian re-identification method based on DCCA fusion features, which is specifically implemented according to the following steps: preprocessing a pedestrian re-identification data set, and adjusting the image size to an appropriate size; and processing the pedestrian data set based on vgg16 The deep convolutional neural network and the omni-scale deep convolutional neural network are used for deep feature extraction respectively; the extracted deep features are subjected to a typical correlation analysis, the respective projection matrices are solved, and the projected features are fused according to the feature fusion strategy; The whole person re-identification process is completed with the fused features. A pedestrian re-identification method based on DCCA fusion features of the present invention, combined with the advantages of vgg16 and omni-scale deep network, improves the robustness of features, effectively eliminates redundant information while fusing features, and improves feature discrimination capability , to improve the accuracy of pedestrian re-identification. |