HMMN: Hard Sample Mining Memory Network for Re-identification

Authors: Qing Li, Pengcheng Han, Shuhui Bu, Ke Li

Abstract: In this paper, we exploit a novel memory-based strategy for mining hard triplets in the field of re-identification. This strategy is realized with an end-to-end deep learning based framework using an external memory pool. It could be used in the security system. We show that the pipeline is able to explicitly provide hard negative and positive samples to generate effective triplets, which are important for online metric learning. That’s because the representative triplets could provide distinctive information to help understand the concept of metric learning between categories. In addition, a ‘focal-triplet loss’ function is proposed to help deal with the lack of positive or negative samples for one anchor, and imbalance between easy and hard triplets for mini-batch. Experimental results on Market-1501, CUHK03 and DukeMTMC-reID demonstrate the effectiveness of our method, and its performance outperforms that of some existing methods.

---

---