Research map / Detection and segmentation / CNN features
WaveFormer: Frequency-Time Decoupled Vision Modeling with Wave Equation
Zishan Shu, Juntong Wu, Wei Yan, Xudong Liu, Hongyu Zhang, Chang Liu, Youdong Mao, Jie Chen
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 |
|---|---|---|---|
| use wave propagation (WaveFormer-T) instead of heat conduction (vHeat-T) | WaveFormer-T · ImageNet-1K val | top-1 accuracy: 82.2 -> 82.5 | Improved |
| use WaveFormer-B instead of Swin-B | WaveFormer-B · ImageNet-1K val | top-1 accuracy: 83.5 -> 84.2 | Improved |
| use WaveFormer-T instead of vHeat-T | WaveFormer-T · COCO box AP (1x) | box AP: 45.1 -> 45.8 | Improved |
| use WaveFormer-T instead of vHeat-T, throughput | WaveFormer-T · COCO Mask R-CNN (1x) | FPS: 32.7 -> 32.1 FPS | Hurt |
| use WaveFormer-B instead of vHeat-B | WaveFormer-B · ADE20K (UperNet) | mIoU: 49.6 -> 50.5 | Improved |
| use WaveFormer-T instead of vHeat-T | WaveFormer-T · COCO box AP (3x) | box AP: 47.2 -> 47.4 | Improved |
| use WaveFormer-S instead of vHeat-S | WaveFormer-S · ImageNet-1K val | top-1 accuracy: 83.6 -> 83.9 | Improved |
| use moderate damping and velocity (v=1, alpha=0.1) instead of extreme values (ablation) | WaveFormer-B · ADE20K (UperNet) | mIoU | Improved |
Closest papers by meaning
- Language-guided Frequency Modulation for Large Vision-Language Models (CVPR 2026 · similarity 0.85)
- Large Spatial Model: End-to-end Unposed Images to Semantic 3D (NeurIPS 2024 · similarity 0.84)
- Large Vision-Language Models Get Lost in Attention (ICML 2026 · similarity 0.84)
- Balancing Frequencies and Pixels in Flow Matching (NeurIPS 2026 · similarity 0.84)
- Unleashing Diffusion Transformers for Visual Correspondence by Modulating Massive Activations (NeurIPS 2025 · similarity 0.84)
- WST: Wavelet-Based Multi-scale Tuning for Visual Transfer Learning (AAAI 2025 · 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.
