Research map / Video / Video segmentation and tracking
WavePolyp: Video Polyp Segmentation via Hierarchical Wavelet-Based Feature Aggregation and Inter-Frame Divergence Perception
Yuhua Zhang, Guilian Chen, Yuanqin He, Huisi Wu, Jing Qin
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 HWFA and IDP to the baseline (ablation) | WavePolyp baseline · SUN-SEG-Hard | Dice: 86.12 -> 87.55 | Improved |
| add HWFA and IDP to the baseline (ablation) | WavePolyp baseline · SUN-SEG-Easy | Dice: 87.51 -> 88.96 | Improved |
| add HFC and LFC without AFA (ablation) | WavePolyp · SUN-SEG-Easy | Sα: 89.12 -> 87.92 | Hurt |
| add AFA alignment when fusing HFC/LFC features (ablation) | WavePolyp · SUN-SEG-Hard | Dice: 85.55 -> 86.76 | Improved |
| start hierarchical fusion at f3 instead of f2 (ablation) | HWFA · SUN-SEG-Hard | Dice: 87.03 -> 87.55 | Improved |
| start hierarchical fusion at f3 instead of applying only to f1 without AFA (ablation) | HWFA · SUN-SEG-Hard | Dice: 85.94 -> 87.55 | Improved |
| start fusion at f4 instead of f3 (ablation) | HWFA · SUN-SEG-Hard | Dice: 87.55 -> 87.47 | No clear effect |
| raise backbone base lr from 1e-5 to 1e-4 (ablation) | WavePolyp · SUN-SEG-Hard | Dice: 87.55 -> 86.97 | Hurt |
| use channel dimension C=64 instead of 32 (ablation) | WavePolyp · SUN-SEG-Hard | Dice: 87.01 -> 87.55 | Improved |
| use channel dimension 256 instead of 64 (ablation) | WavePolyp · SUN-SEG-Hard | Dice: 87.55 -> 87.44 | No clear effect |
| use clip length 5 rather than shorter or longer clips (ablation) | WavePolyp · SUN-SEG | Dice | Improved |
| use WavePolyp instead of VP-SAM | VPS methods · SUN-SEG-Hard | Dice: 86.94 -> 87.55 | Improved |
Closest papers by meaning
- An Embedding-Unleashing Video Polyp Segmentation Framework via Region Linking and Scale Alignment (AAAI 2024 · similarity 0.88)
- Polyper: Boundary Sensitive Polyp Segmentation (AAAI 2024 · similarity 0.86)
- Boost the Inference with Co-training: A Depth-guided Mutual Learning Framework for Semi-supervised Medical Polyp Segmentation (CVPR 2025 · similarity 0.83)
- beta-FFT: Nonlinear Interpolation and Differentiated Training Strategies for Semi-Supervised Medical Image Segmentation (CVPR 2025 · similarity 0.82)
- Complex Wavelet Mutual Information Loss: A Multi-Scale Loss Function for Semantic Segmentation (ICML 2025 · similarity 0.82)
- Multi-view Masked Contrastive Representation Learning for Endoscopic Video Analysis (NeurIPS 2024 · similarity 0.81)
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.
