Context (archaeology)

Linguistics

Philosophy

Smoothing

Biology

Discriminative model

Identification (biology)

Feature learning

Similarity (geometry)

Exploit

Computer security

Operating system

Economics

Set (abstract data type)

Data mining

Computer science

Botany

Code (set theory)

Paleontology

Task (project management)

Source code

Pattern recognition (psychology)

Management

Image (mathematics)

Feature (linguistics)

Artificial intelligence

Cluster analysis

Programming language

Noise (video)

Machine learning

Computer vision

Unsupervised person re-identification (re-ID) aims at learning discriminative representations for person retrieval from unlabeled data. Recent techniques accomplish this task by using pseudo-labels, but these labels are inherently noisy and deteriorate the accuracy. To overcome this problem, several pseudo-label refinement methods have been proposed, but they neglect the fine-grained local context essential for person re-ID. In this paper, we propose a novel Part-based Pseudo Label Refinement (PPLR) framework that reduces the label noise by employing the complementary relationship between global and part features. Specifically, we design a cross agreement score as the similarity of k-nearest neighbors between feature spaces to exploit the reliable complementary relationship. Based on the cross agreement, we refine pseudo-labels of global features by ensembling the predictions of part features, which collectively alleviate the noise in global feature clustering. We further refine pseudo-labels of part features by applying label smoothing according to the suitability of given labels for each part. Thanks to the reliable complementary information provided by the cross agreement score, our PPLR effectively reduces the influence of noisy labels and learns discriminative representations with rich local contexts. Extensive experimental results on Market-1501 and MSMT17 demonstrate the effectiveness of the proposed method over the state-of-the-art performance. The code is available at https://github.com/yoonkicho/PPLR.

Part-based Pseudo Label Refinement for Unsupervised Person Re-identification