Fine-Grained Classification with Noisy Labels

TL;DR

This paper introduces SNSCL, a novel contrastive learning framework designed to improve fine-grained classification accuracy in noisy label scenarios, addressing the limitations of existing methods.

Contribution

It proposes a noise-tolerant contrastive learning approach with a weight-aware mechanism and stochastic sampling, specifically tailored for fine-grained noisy label datasets.

Findings

SNSCL outperforms existing methods on fine-grained noisy datasets.

The proposed framework effectively mitigates label noise impacts.

Experimental results demonstrate improved classification accuracy.

Abstract

Learning with noisy labels (LNL) aims to ensure model generalization given a label-corrupted training set. In this work, we investigate a rarely studied scenario of LNL on fine-grained datasets (LNL-FG), which is more practical and challenging as large inter-class ambiguities among fine-grained classes cause more noisy labels. We empirically show that existing methods that work well for LNL fail to achieve satisfying performance for LNL-FG, arising the practical need of effective solutions for LNL-FG. To this end, we propose a novel framework called stochastic noise-tolerated supervised contrastive learning (SNSCL) that confronts label noise by encouraging distinguishable representation. Specifically, we design a noise-tolerated supervised contrastive learning loss that incorporates a weight-aware mechanism for noisy label correction and selectively updating momentum queue lists. By this mechanism, we mitigate the effects of noisy anchors and avoid inserting noisy labels into the momentum-updated queue. Besides, to avoid manually-defined augmentation strategies in contrastive learning, we propose an efficient stochastic module that samples feature embeddings from a generated distribution, which can also enhance the representation ability of deep models. SNSCL is general and compatible with prevailing robust LNL strategies to improve their performance for LNL-FG. Extensive experiments demonstrate the effectiveness of SNSCL.