Research map / Detection and segmentation / CNN features
CycleMLP: A MLP-Like Architecture for Dense Visual Predictions
Shoufa Chen, Enze Xie, Chongjian Ge, Runjian Chen, Ding Liang, Ping Luo
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 Cycle FC instead of Channel FC for spatial mixing (ablation) | CycleMLP-B2 · ADE20K val | mIoU: 36.3 -> 42.4 | Improved |
| use Cycle FC instead of Channel FC for spatial mixing (ablation) | CycleMLP-B2 · COCO val2017 | AP: 35.0 -> 41.7 | Improved |
| use Cycle FC instead of Channel FC for spatial mixing (ablation) | CycleMLP-B2 · ImageNet-1K val | top-1 accuracy: 79.4 -> 81.6 | Improved |
| use Cycle FC instead of Spatial FC (ablation) | CycleMLP-B2 · ImageNet-1K val | top-1 accuracy: 80.9 -> 81.6 | Improved |
| add the 7x1 branch to the 1x7 plus 1x1 block (ablation) | CycleMLP-B2 · ImageNet-1K val | top-1 accuracy: 80.5 -> 81.6 | Improved |
| add the 1x7 branch to the 7x1 plus 1x1 block (ablation) | CycleMLP-B2 · ImageNet-1K val | top-1 accuracy: 80.4 -> 81.6 | Improved |
| add the 1x1 branch to the 1x7 plus 7x1 block (ablation) | CycleMLP-B2 · ImageNet-1K val | top-1 accuracy: 81.3 -> 81.6 | Improved |
| use distinct 1x7 and 7x1 branches rather than duplicating a branch at matched parameters (ablation) | CycleMLP-B2 · ImageNet-1K val | top-1 accuracy: 80.6 -> 81.6 | Improved |
| increase stepsize from 3 to 7 (ablation) | CycleMLP-B2 · ADE20K val | mIoU: 42.4 -> 43.9 | Improved |
| increase stepsize from 3 to 9 (ablation) | CycleMLP-B2 · ImageNet-1K val | top-1 accuracy: 81.6 -> 81.5 | No clear effect |
| replace cyclical sampling with random sampling (ablation) | CycleMLP · ImageNet-1K val | top-1 accuracy: averaged Top-1 accuracy drops by 1.3% | Hurt |
| replace cyclical sampling with dilated sampling (ablation) | CycleMLP · ImageNet-1K val | top-1 accuracy: -0.5% acc | Hurt |
Closest papers by meaning
- CycleMLP: A MLP-like Architecture for Dense Prediction (ICLR 2022 · similarity 0.99)
- MorphMLP: An Efficient MLP-Like Backbone for Spatial-Temporal Representation Learning (ECCV 2022 · similarity 0.86)
- Vision Permutator: A Permutable MLP-Like Architecture for Visual Recognition (TPAMI 2022 · similarity 0.85)
- RepMLPNet: Hierarchical Vision MLP With Re-Parameterized Locality (CVPR 2022 · similarity 0.85)
- SOLO: A Simple Framework for Instance Segmentation (TPAMI 2021 · similarity 0.85)
- RMP-SAM: Towards Real-Time Multi-Purpose Segment Anything (ICLR 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.
