Unsupervised Person Re-identification with Stochastic Training Strategy
This work addresses camera variance and clustering errors in unsupervised person re-identification, which is important for scalable surveillance applications, but it is incremental as it builds on existing clustering-based methods.
The paper tackles the problem of unreliable pseudo labels and feature inconsistency in unsupervised person re-identification by proposing a stochastic training strategy that uses random instances from clusters to update memory for contrastive learning, achieving state-of-the-art results with 85.8% mAP on Market-1501 and 73.2% mAP on DukeMTMC-reID.
Unsupervised person re-identification (re-ID) has attracted increasing research interests because of its scalability and possibility for real-world applications. State-of-the-art unsupervised re-ID methods usually follow a clustering-based strategy, which generates pseudo labels by clustering and maintains a memory to store instance features and represent the centroid of the clusters for contrastive learning. This approach suffers two problems. First, the centroid generated by unsupervised learning may not be a perfect prototype. Forcing images to get closer to the centroid emphasizes the result of clustering, which could accumulate clustering errors during iterations. Second, previous methods utilize features obtained at different training iterations to represent one centroid, which is not consistent with the current training sample, since the features are not directly comparable. To this end, we propose an unsupervised re-ID approach with a stochastic learning strategy. Specifically, we adopt a stochastic updated memory, where a random instance from a cluster is used to update the cluster-level memory for contrastive learning. In this way, the relationship between randomly selected pair of images are learned to avoid the training bias caused by unreliable pseudo labels. The stochastic memory is also always up-to-date for classifying to keep the consistency. Besides, to relieve the issue of camera variance, a unified distance matrix is proposed during clustering, where the distance bias from different camera domain is reduced and the variances of identities is emphasized.