Research map / Vision-language and multimodal / CLIP and zero-shot
Bridging the Pathology Domain Gap: Efficiently Adapting CLIP for Pathology Image Analysis with Limited Labeled Data
Zhengfeng Lai, Joohi Chauhan, Brittany N. Dugger, Chen-Nee Chuah
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 RFR (residual feature refinement) to frozen CLIP (ablation) | CLIP ResNet-50 · PCam (0.5% labeled) | accuracy: 56.5 -> 72.0 | Improved |
| add Hidden Representation Perturbation (HRP) (ablation) | Path-CLIP · PCam (0.5% labeled) | accuracy: 72.0 -> 74.2 | Improved |
| add Dual-view Vision Contrastive (DVC) bound constraints for residual ratio (ablation) | Path-CLIP · PCam (0.5% labeled) | accuracy: 74.2 -> 78.5 | Improved |
| add Doublet Multimodal Contrastive Loss (DMCL) (ablation) | Path-CLIP · PCam (0.5% labeled) | accuracy: 78.5 -> 81.9 | Improved |
| fully fine-tune CLIP on 1% PCam instead of zero-shot | CLIP ResNet-50 · PCam (1% labeled) | accuracy: 56.5% -> 40.3% | Hurt |
| lower learning rate for full fine-tuning (100 epochs) | CLIP ResNet-50 · PCam (1% labeled) | accuracy: 35.9% -> 45.6% | Mixed |
| increase labeled data for CoOp prompt tuning | CoOp · PCam | accuracy: 63.6 -> 61.5 (0.5% to 1.5%) | No clear effect |
| Path-CLIP versus TIP-Adapter on PCam | CLIP · PCam (0.5% labeled) | accuracy: 73.9 -> 81.9 | Improved |
| Path-CLIP versus supervised U-Net at low data | GM/WM segmentation · GM/WM segmentation (0.1% data) | GM IoU: 46.13 -> 78.05 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- TIP-Adapter — Tip-Adapter: Training-Free Adaption of CLIP for Few-Shot Classification (ECCV 2022) · better on 4 of 4 shared results
- CLIP-Adapter — CLIP-Adapter: Better Vision-Language Models with Feature Adapters (IJCV 2023) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- PathFLIP: Fine-grained Language-Image Pretraining for Versatile Computational Pathology (AAAI 2026 · similarity 0.89)
- PathGen-1.6M: 1.6 Million Pathology Image-text Pairs Generation through Multi-agent Collaboration (ICLR 2025 · similarity 0.89)
- PathView-Bench: Can Multimodal Large Language Models Achieve Fine-grained Multiscale Understanding of Pathology Images? (NeurIPS 2026 · similarity 0.89)
- CPath-Omni: A Unified Multimodal Foundation Model for Patch and Whole Slide Image Analysis in Computational Pathology (CVPR 2025 · similarity 0.89)
- PathVQ: Reforming Computational Pathology Foundation Model for Whole Slide Image Analysis via Vector Quantization (NeurIPS 2025 · similarity 0.88)
- FOCUS: Knowledge-enhanced Adaptive Visual Compression for Few-shot Whole Slide Image Classification (CVPR 2025 · similarity 0.87)
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.
