Establish, Break, Rebuild, Consolidate: Relational Self-Supervised Learning for Cross-Dataset Skin-Lesion Representation Learning

Skin-lesion representations should preserve clinically meaningful relationships across acquisition conditions, yet dermoscopic and clinical or smartphone images can differ substantially in illumination, scale, viewpoint, color response, and visible structure. We study self-supervised representation learning from independently collected and unpaired datasets without using diagnosis labels to define cross-dataset correspondences. We introduce Cross-dataset INter-view relational Distillation with Evidence-masked Recovery and Stabilization (CINDERS), which learns representations by explicitly modeling relationships among samples. An exponential-moving-average teacher defines the relational organization of intact multi-view images and ranks local visual evidence. A shared student first establishes this intact relation; the highest-ranked patches are then physically removed before a second frozen backbone pass; the student rebuilds the intact relation from the surviving evidence; and a symmetric contrastive objective consolidates the intact and rebuilt representations. We evaluate the learned representation primarily through cross-acquisition retrieval at rank 1 (R@1). On DermaMNIST-C and PAD-UFES-20, CINDERS improves the two cross-acquisition R@1 directions from $0.2665/0.2799$ for frozen DINOv3 to $0.5330/0.4940$, increasing their harmonic mean from $0.2730$ to $0.5128$. Under the matched self-supervised learning protocol, CINDERS achieves a four-direction mean R@1 of $0.6014$, compared with $0.5527$ for iBOT, while its cross-acquisition harmonic mean of $0.5128$ exceeds the $0.3716$ obtained by Relational Knowledge Distillation. On the external Derm7pt–MILK10K evaluation, the four-direction mean R@1 increases from $0.2961$ for frozen DINOv3 to $0.3665$, and the cross-acquisition harmonic mean increases from $0.2638$ to $0.3621$. Together, the internal and external results show that recovering inter-sample relationships from incomplete visual evidence can improve cross-acquisition representation compatibility without diagnosis supervision or manually specified cross-dataset positive pairs.