Research map / Vision-language and multimodal / Cross-modal learning
Pairing-free Group-level Knowledge Distillation for Robust Gastrointestinal Lesion Classification in White-Light Endoscopy
Qiang Hu, Qimei Wang, Yingjie Guo, Qiang Li, Zhiwei Wang
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 GKD-Pro and GKD-Den to a plain ResNet-50 classifier (ablation) | WLI classifier · PICCOLO | AUC: 71.2 -> 90.1 | Improved |
| add GKD-Pro alone (ablation) | WLI classifier · PICCOLO | AUC: 71.2 -> 83.5 | Improved |
| add GKD-Den alone (ablation) | WLI classifier · PICCOLO | AUC: 71.2 -> 85.0 | Improved |
| add GKD-Den to GKD-Pro (ablation) | PaGKD · IH-GC | AUC: 75.5 -> 84.0 | Improved |
| use LR-QFormer prototypes instead of pooled features (ablation) | GKD-Pro · PICCOLO | AUC: 85.7 -> 90.1 | Improved |
| use semantic relation-guided cross-attention instead of standard cross-attention (ablation) | GKD-Den · PICCOLO | AUC: 83.2 -> 90.1 | Improved |
| add the WLI-to-NBI direction of the consistency loss (ablation) | GKD-Den · IH-GC | AUC: 81.9 -> 84.0 | Improved |
| use group-level rather than image-level distillation for GKD-Pro (ablation) | GKD-Pro · PICCOLO | AUC: 78.6 -> 83.5 | Improved |
| use group-level rather than image-level distillation for both modules (ablation) | PaGKD · IH-GC | AUC: 75.7 -> 84.0 | Improved |
| add unpaired training data to ADD (ablation) | ADD · IH-Polyp | AUC: 82.8 -> 75.7 | Hurt |
| add unpaired training data to PolypsAlign (ablation) | PolypsAlign · IH-Polyp | AUC: 74.3 -> 60.6 | Hurt |
| add unpaired training data to PaGKD (ablation) | PaGKD · PICCOLO | AUC: 84.4 -> 90.1 | Improved |
Closest papers by meaning
- Gastric-X: A Multimodal Multi-Phase Benchmark Dataset for Advancing Vision-Language Models in Gastric Cancer Analysis (CVPR 2026 · similarity 0.85)
- Clinical Cognition Alignment for Gastrointestinal Diagnosis with Multimodal LLMs (ECCV 2026 · similarity 0.85)
- What Can Be Transferred: Unsupervised Domain Adaptation for Endoscopic Lesions Segmentation (CVPR 2020 · similarity 0.85)
- Prompt as Knowledge Bank: Boost Vision-language model via Structural Representation for zero-shot medical detection (ICLR 2025 · similarity 0.84)
- FOCUS: Knowledge-enhanced Adaptive Visual Compression for Few-shot Whole Slide Image Classification (CVPR 2025 · similarity 0.83)
- EndoBench: A Comprehensive Evaluation of Multi-Modal Large Language Models for Endoscopy Analysis (NeurIPS 2025 · similarity 0.83)
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.
