mlsanityData to 14 Oct 2026

TopoPult-SSL: Gland-Mask-Free Cross-Device Meibomian Gland Segmentation via Self-Distilled Weak Clinical Priors

arXiv 2606.05347 ↗First posted 3 Jun 2026cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

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.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 8 of its 19 references that are in the map. First posted on arXiv 3 Jun 2026.

Builds on

8 of 19 references.

References · 8

  1. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
  2. 2015Distilling the Knowledge in a Neural Networkcited by 10,324
  3. 2023Segment Anythingcited by 8,234
  4. 2020FixMatch: Simplifying Semi-Supervised Learning with Consistency and Confidencecited by 2,085
  5. 2021Semi-Supervised Semantic Segmentation with Cross Pseudo Supervisioncited by 463
  6. 2019Uncertainty-aware Self-ensembling Model for Semi-supervised 3D Left Atrium Segmentationcited by 379
  7. 2020clDice -- A Novel Topology-Preserving Loss Function for Tubular Structure Segmentationcited by 294
  8. 2018Unsupervised domain adaptation for medical imaging segmentation with self-ensemblingcited by 94

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Weakly Supervised Teacher-Student Framework with Progressive Pseudo-mask Refinement for Gland Segmentation
  2. 2024Mask-guided cross-image attention for zero-shot in-silico histopathologic image generation with a diffusion model
  3. 2024EyeDiff: text-to-image diffusion model improves rare eye disease diagnosis
  4. 2025Robust Multimodal Learning for Ophthalmic Disease Grading via Disentangled Representation
  5. 2024Masked Attention as a Mechanism for Improving Interpretability of Vision Transformers
  6. 2026FOCUS: Benchmarking Retinal Model Generalization from Foundation Vision Encoders to Multimodal LLMs
  7. 2024Synthetic Forehead-creases Biometric Generation for Reliable User Verification
  8. 2025LMOD+: A Comprehensive Multimodal Dataset and Benchmark for Developing and Evaluating Multimodal Large Language Models in Ophthalmology