Search papers, labs, and topics across Lattice.
This paper introduces Recurrent Contrastive Learning (RCL) to address class imbalance in medical image classification by expanding the latent support region of tail classes. By utilizing DINOv3 with LoRA adapters and a Temporal Memory Queue (TMQ) to preserve historical feature states, RCL effectively enhances the representation of tail classes and mitigates encroachment from head classes. Experimental results across three imbalanced medical datasets show that RCL consistently outperforms existing methods, leading to improved inter-class separation and classification accuracy.
Tail classes in medical image classification can be effectively shielded from head-class encroachment, leading to significant improvements in performance.
Medical image classification often suffers from class imbalance due to the inherent disparities in disease incidence. Existing approaches, such as class resampling and loss reweighting, mainly improve learning within the observed feature distribution, but do not explicitly enlarge the latent support region of tail classes. As a result, tail-class representations remain overly compact and are easily encroached upon by head classes, leading to biased decision boundaries. In this work, we propose Recurrent Contrastive Learning (RCL) for imbalanced medical image classification. RCL progressively expands the support region of tail classes by recurrently reusing historical feature states across training phases. Specifically, we adopt DINOv3 with LoRA adapters as the backbone to provide robust feature embeddings. We then devise a Temporal Memory Queue (TMQ) to preserve corpus-level features across training phases and provide diversified global references for contrastive learning. Based on TMQ, we construct Temporal Anchors (TARs) to form an anchor field around tail classes. This field enlarges the support region of tail classes, suppresses head-class encroachment, and improves inter-class separation. Extensive experiments on three imbalanced medical datasets demonstrate that RCL achieves consistent improvements over strong baselines. The code is available at https://github.com/dndins/RCL.
