HBReID: Harder Batch for Re-identification
This work addresses performance limitations in person re-identification for applications like surveillance, though it is incremental as it builds on existing triplet loss methods.
The paper tackles the problem of triplet loss in person re-identification by proposing a hard batch mining method to globally select the hardest samples and an adversarial scene removal module for scene-invariant features, achieving state-of-the-art results on the MSMT17 dataset.
Triplet loss is a widely adopted loss function in ReID task which pulls the hardest positive pairs close and pushes the hardest negative pairs far away. However, the selected samples are not the hardest globally, but the hardest only in a mini-batch, which will affect the performance. In this report, a hard batch mining method is proposed to mine the hardest samples globally to make triplet harder. More specifically, the most similar classes are selected into a same mini-batch so that the similar classes could be pushed further away. Besides, an adversarial scene removal module composed of a scene classifier and an adversarial loss is used to learn scene invariant feature representations. Experiments are conducted on dataset MSMT17 to prove the effectiveness, and our method surpasses all of the previous methods and sets state-of-the-art result.