SphereReID: Deep Hypersphere Manifold Embedding for Person Re-Identification
This work addresses person re-identification for surveillance and security applications, offering an incremental improvement over existing methods.
The paper tackles the problem of person re-identification by proposing SphereReID, a method that uses a modified softmax function to learn hypersphere manifold embeddings, achieving state-of-the-art results such as 94.4% rank-1 accuracy on Market-1501 and 83.9% on DukeMTMC-reID.
Many current successful Person Re-Identification(ReID) methods train a model with the softmax loss function to classify images of different persons and obtain the feature vectors at the same time. However, the underlying feature embedding space is ignored. In this paper, we use a modified softmax function, termed Sphere Softmax, to solve the classification problem and learn a hypersphere manifold embedding simultaneously. A balanced sampling strategy is also introduced. Finally, we propose a convolutional neural network called SphereReID adopting Sphere Softmax and training a single model end-to-end with a new warming-up learning rate schedule on four challenging datasets including Market-1501, DukeMTMC-reID, CHHK-03, and CUHK-SYSU. Experimental results demonstrate that this single model outperforms the state-of-the-art methods on all four datasets without fine-tuning or re-ranking. For example, it achieves 94.4% rank-1 accuracy on Market-1501 and 83.9% rank-1 accuracy on DukeMTMC-reID. The code and trained weights of our model will be released.