TopoPult-SSL: Gland-Mask-Free Cross-Device Meibomian Gland Segmentation via Self-Distilled Weak Clinical Priors
For ophthalmologists and medical imaging researchers, this reduces annotation cost by leveraging routinely recorded clinical signals instead of expensive gland masks for domain adaptation.
TopoPult-SSL enables cross-device meibomian gland segmentation without target gland masks, using weak clinical priors (eyelid outlines, Pult grades). It achieves Dice 0.716 on the MGD-1k to CAMG benchmark, outperforming UA-MT (0.710) and ensemble teacher (0.720), and its gland-mask-free variant achieves Precision 0.694 vs. 0.30-0.34 for SAM/MedSAM.
Every new clinical imaging device creates a domain shift where dense gland masks are expensive yet cheap clinical signals -- eyelid outlines, Pult grades, morphometric ratios -- are routinely recorded. We present TopoPult-SSL, a two-stage framework for cross-device meibomian gland segmentation. Stage 1 adapts a source-trained model without target gland masks in the training loss, using four weak-prior anchors driven by target eyelid masks and clinical metadata only. Stage 2, when target gland masks are available, distils complementary Stage-1 teachers into a single compact student via supervised self-distillation. We develop and validate the technique on the public MGD-1k to CAMG research benchmark (1,000 to 100 images, different device), where the distilled model achieves Dice 0.716+/-0.006 (best 0.726), surpassing UA-MT (0.710) and the ensemble teacher (0.720) -- with a single pass. The gland-mask-free Stage-1 variant reaches Precision 0.694 vs. 0.30-0.34 for SAM/MedSAM (p<0.001), enabling deployment without dense gland contouring. Code and reproducibility scripts are released.