Mirror-Fusion Attention for Reflection-Aware Self-Supervised Representation Learning
For practitioners working with approximately symmetric data (medical images, faces), MFASSL offers a simple, parameter-efficient way to improve SSL performance and robustness without redesigning backbones.
MFASSL improves self-supervised learning on bilateral data by injecting a soft reflection prior via a lightweight Mirror-Fusion Attention module, achieving consistent gains over MoCo-v3, DINO, and MAE on CheXpert, BraTS, CelebA-HQ, and WFLW with only ~2.7% additional parameters.
Most self-supervised learning (SSL) methods encourage invariance across augmentations, but strict flip invariance can suppress informative left--right correspondences in approximately bilateral data such as medical images and human faces. We propose Mirror-Fusion-Augmented Self-Supervised Learning (MFASSL), a Vision Transformer framework that injects a soft reflection prior into standard SSL without redesigning the backbone. MFASSL constructs mirror-paired views aligned to an estimated symmetry axis and introduces a lightweight Mirror-Fusion Attention (MFA) module for adaptive token-level interaction between mirrored regions while preserving asymmetric cues. The base SSL objective is further coupled with reflection-consistency and mid-layer token-alignment losses. Across CheXpert, BraTS, CelebA-HQ, and WFLW, MFASSL improves downstream performance, calibration, and reflection robustness over MoCo-v3, DINO, and MAE baselines under matched ViT-B/16 settings. It also achieves stronger and more consistent gains than recent equivariant SSL approaches with only approximately 2.7\% additional parameters. These results show that lightweight geometry-aware priors can effectively complement invariance-based SSL.