Circle-Based Ratio Loss for Person Reidentification

Figure 3

The network architecture for our person reidentification experiments is comprised of the backbone, global average pooling layer, batch normalization layer, fully connected layer, and L2 normalization layer. In the training procedure, the training images are organized as PK format in which P and K denote the number of identities and the sample number for each identity, respectively. Then the model learns the pedestrian features under the supervision of the ID loss and the ratio loss. In the testing phase, the last fully connected layer is removed and the remaining networks make up the feature extractor. The testing images are fed to the feature extractor to obtain pedestrian features, and the re-id task is conducted by comparing the similarity between extracted features.