Research map / Label-efficient and robust learning / Recognition and long-tail
Homeomorphism Prior for False Positive and Negative Problem in Medical Image Dense Contrastive Representation Learning
Yuting He, Boyu Wang, Rongjun Ge, Yang Chen, Guanyu Yang, Shuo Li
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add GEMINI dense contrastive pre-learning to a U-Net trained with five labels | U-Net · FS-Semi average | DSC: 79.1 -> 88.7 | Improved |
| replace GVSL-Semi with GEMINI-Semi (add GSS, smooth, bijection) | GVSL-Semi · FS-Semi average | DSC: 86.2 -> 88.7 | Improved |
| add GEMINI to a TransUNet backbone | TransUNet · FS-Semi average | DSC: 72.8 -> 87.9 | Improved |
| add GEMINI to a SwinUNet backbone | SwinUNet · FS-Semi average | DSC: 57.5 -> 84.9 | Improved |
| add GEMINI-MIP pre-training instead of random initialization | 2D U-Net · SS-MIP average | DSC: 73.6 -> 87.0 | Improved |
| add GSS and smoothness on top of GVSL-MIP | GVSL-MIP · SS-MIP average | DSC: 86.1 -> 87.0 | Improved |
| add the geometric visual similarity (GVS) alone for correspondence learning (ablation) | GEMINI-Semi · T3 2D chest structures | DSC | Hurt |
| add the geometric semantic similarity (GSS) to GVS (ablation) | GEMINI-Semi · T1 and T3 | DSC | Improved |
| add smoothness loss for continuity (ablation) | GEMINI-Semi · T1 and T3 | DSC | Improved |
| add the bijection (inverse mapping) loss (ablation) | GEMINI-Semi · T1 and T3 | DSC | Improved |
| enlarge the Deformer receptive field r (ablation) | GEMINI-Semi · T1 3D cardiac | DSC: around 90% DSC | No clear effect |
| add the DHL deformer module (soft learning of feature pairs) (ablation) | GEMINI-Semi · T1 3D cardiac | DSC: more than 5% DSC | Improved |
Closest papers by meaning
- TopoCL: Topological Contrastive Learning for Medical Imaging (CVPR 2026 · similarity 0.86)
- G2D: From Global to Dense Radiography Representation Learning via Vision-Language Pre-training (NeurIPS 2024 · similarity 0.85)
- Geometric Visual Similarity Learning in 3D Medical Image Self-Supervised Pre-Training (CVPR 2023 · similarity 0.85)
- Neighbor Does Matter: Density-Aware Contrastive Learning for Medical Semi-supervised Segmentation (AAAI 2025 · similarity 0.84)
- Autoregressive Sequence Modeling for 3D Medical Image Representation (AAAI 2025 · similarity 0.84)
- Diffusion Model for Dense Matching (ICLR 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
