Deep adaptive feature embedding with local sample distributions for person re-identification

- An improved deep feature embedding approach for person re-identification is presented to learn representations amenable to similarity score computation.
- The quality of learned representations and the training efficiency is augmented by jointly optimizing robust feature embedding, local adaptive similarity learning, and suitable positive mining.
- An alternative to CNN embedding is presented by formulating a stacked CRBMs into local sample structure in deep feature space, and thus enables local adaptive similarity metric learning as well as plausible positive mining.
- Stochastic gradient descent is modified to reuse past computed gradients from neighborhood data points, leading to linear convergence.

Person re-identification (re-id) aims to match pedestrians observed by disjoint camera views. It attracts increasing attention in computer vision due to its importance to surveillance systems. To combat the major challenge of cross-view visual variations, deep embedding approaches are proposed by learning a compact feature space from images such that the Euclidean distances correspond to their cross-view similarity metric. However, the global Euclidean distance cannot faithfully characterize the ideal similarity in a complex visual feature space because features of pedestrian images exhibit unknown distributions due to large variations in poses, illumination and occlusion. Moreover, intra-personal training samples within a local range which are robust to guide deep embedding against uncontrolled variations cannot be captured by a global Euclidean distance. In this paper, we study the problem of person re-id by proposing a novel sampling to mine suitable positives (i.e., intra-class) within a local range to improve the deep embedding in the context of large intra-class variations. Our method is capable of learning a deep similarity metric adaptive to local sample structure by minimizing each sample's local distances while propagating through the relationship between samples to attain the whole intra-class minimization. To this end, a novel objective function is proposed to jointly optimize similarity metric learning, local positive mining and robust deep feature embedding. This attains local discriminations by selecting local-ranged positive samples, and the learned features are robust to dramatic intra-class variations. Experiments on benchmarks show state-of-the-art results achieved by our method.