Research map / Detection and segmentation / Weakly supervised detection and segmentation
MCTformer+: Multi-Class Token Transformer for Weakly Supervised Semantic Segmentation
Lian Xu, Mohammed Bennamoun, Farid Boussaïd, Hamid Laga, Wanli Ouyang, Dan Xu
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 multiple class tokens (MCTformer attention) instead of a single class token with CAM (TS-CAM*) (ablation) | MCTformer · PASCAL VOC train (seed) | mIoU: 41.3 -> 47.2 | Improved |
| add PatchAffinity refinement to attention maps (ablation) | MCTformer (Attention) · PASCAL VOC train (seed) | mIoU: 47.2 -> 55.2 | Improved |
| add PatchCAM fusion to attention maps (ablation) | MCTformer (Attention) · PASCAL VOC train (seed) | mIoU: 47.2 -> 58.2 | Improved |
| add PatchAffinity refinement to the fused PatchCAM and attention maps (ablation) | MCTformer (Attention + PatchCAM) · PASCAL VOC train (seed) | mIoU: 58.2 -> 61.7 | Improved |
| add PatchAffinity to maps for training the segmentation network (ablation) | MCTformer · PASCAL VOC val (segmentation) | mIoU: 55.6 -> 58.8; 61.1 -> 62.6 | Improved |
| average pooling of class tokens instead of a fully-connected layer for class prediction (ablation) | MCTformer · PASCAL VOC train (seed) | mIoU: 41.5 -> 47.2 | Improved |
| average pooling of class tokens instead of max pooling (ablation) | MCTformer · PASCAL VOC train (seed) | mIoU: 26.8 -> 47.2 | Improved |
| use GWRP instead of GAP in CAM module (ablation) | MCTformer+ · PASCAL VOC train (seed) | mIoU: 66.1 -> 68.8 | Improved |
| use GWRP instead of GMP in CAM module (ablation) | MCTformer+ · PASCAL VOC train (seed) | mIoU: 59.4 -> 68.8 | Improved |
| add CCT module on top of GAP (224x224 and 448x448 training) (ablation) | MCTformer+ · PASCAL VOC train (seed) | mIoU: 63.2 -> 64.5 (224); 63.6 -> 66.1 (448) | Improved |
| add CCT module on top of GWRP (224x224 and 448x448 training) (ablation) | MCTformer+ · PASCAL VOC train (seed) | mIoU: 64.8 -> 65.8 (224); 65.9 -> 68.8 (448) | Improved |
| apply CCT only on last layer instead of no CCT (ablation) | MCTformer+ · PASCAL VOC train (seed) | mIoU: 64.8 | Hurt |
Later papers that compared with this method
6 later accepted papers report MCTformer+ as a baseline in their results tables; 6 of them report a better number on most of the results they share with it.
- Class Token as Proxy: Optimal Transport-assisted Proxy Learning for Weakly Supervised Semantic Segmentation (ICCV 2025 · better on 5 of 5 shared results)
- Bias-Resilient Weakly Supervised Semantic Segmentation Using Normalizing Flows (ICCV 2025 · better on 3 of 3 shared results)
- Frozen CLIP-DINO: A Strong Backbone for Weakly Supervised Semantic Segmentation (TPAMI 2025 · better on 1 of 1 shared result)
- Progressive Feature Self-Reinforcement for Weakly Supervised Semantic Segmentation (AAAI 2024 · better on 3 of 3 shared results)
- Diffusion-Guided Weakly Supervised Semantic Segmentation (ECCV 2024 · better on 2 of 2 shared results)
- Phase Concentration and Shortcut Suppression for Weakly Supervised Semantic Segmentation (ECCV 2024 · 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
- Multi-Class Token Transformer for Weakly Supervised Semantic Segmentation (CVPR 2022 · similarity 0.98)
- Learning Multi-Modal Class-Specific Tokens for Weakly Supervised Dense Object Localization (CVPR 2023 · similarity 0.93)
- Class Tokens Infusion for Weakly Supervised Semantic Segmentation (CVPR 2024 · similarity 0.91)
- Know Your Attention Maps: Class-specific Token Masking for Weakly Supervised Semantic Segmentation (ICCV 2025 · similarity 0.91)
- Token Contrast for Weakly-Supervised Semantic Segmentation (CVPR 2023 · similarity 0.91)
- MoRe: Class Patch Attention Needs Regularization for Weakly Supervised Semantic Segmentation (AAAI 2025 · similarity 0.90)
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.
