Research map / Detection and segmentation / Open-set and few-shot detection
TagCLIP: Improving Discrimination Ability of Zero-Shot Semantic Segmentation
Jingyao Li, Pengguang Chen, Shengju Qian, Shu Liu, Jiaya Jia
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 |
|---|---|---|---|
| insert untrained trusty token directly into baseline (ablation) | one-stage baseline · PASCAL VOC 2012 | mIoU(U): 75.5 -> 43.0 | Hurt |
| add Trusty Learner (attention block) to trusty token (ablation) | baseline + trusty token · PASCAL VOC 2012 | mIoU(U): 43.0 -> 84.9 | Improved |
| add Trusty Learner with trusty token to baseline (ablation) | one-stage baseline · PASCAL VOC 2012 | hIoU: 83.2 -> 89.0 | Improved |
| add weighted map (ablation) | TagCLIP without weighted map · PASCAL VOC 2012 | hIoU: 89.0 -> 89.2 | Improved |
| initialize trusty token with zeros rather than random (ablation) | TagCLIP · PASCAL VOC 2012 | hIoU: 86.7 -> 89.2 | Improved |
| use similarity matrix concatenated with text tokens as attention input instead of other representations (ablation) | Trusty Learner · PASCAL VOC 2012 | mIoU(U): 41.5 -> 84.9 | Improved |
| simplify segmentor (single attention instead of two) | TagCLIP vs ZegCLIP · VOC FPS | FPS: 9.0 -> 13.3 | Improved |
| add self-training (transductive) to TagCLIP | TagCLIP · PASCAL Context | mIoU(U): 65.8 vs ZegCLIP+ST 68.5 | Hurt |
| use TagCLIP instead of ZegCLIP in background-undifferentiated setting | ZegCLIP · COCO-Stuff 164K | mIoU(U): 30.4 -> 36.0 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- ZegCLIP — ZegCLIP: Towards Adapting CLIP for Zero-Shot Semantic Segmentation (CVPR 2023) · better on 8 of 8 shared results
- FreeSeg — FreeSeg: Unified, Universal and Open-Vocabulary Image Segmentation (CVPR 2023) · 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
- SuperCLIP: CLIP with Simple Classification Supervision (NeurIPS 2025 · similarity 0.91)
- Interpreting and Analysing CLIP's Zero-Shot Image Classification via Mutual Knowledge (NeurIPS 2024 · similarity 0.90)
- Cascade-CLIP: Cascaded Vision-Language Embeddings Alignment for Zero-Shot Semantic Segmentation (ICML 2024 · similarity 0.90)
- TagCLIP: A Local-to-Global Framework to Enhance Open-Vocabulary Multi-Label Classification of CLIP without Training (AAAI 2024 · similarity 0.89)
- Exploring Regional Clues in CLIP for Zero-Shot Semantic Segmentation (CVPR 2024 · similarity 0.89)
- Towards Open-Vocabulary Semantic Segmentation Without Semantic Labels (NeurIPS 2024 · similarity 0.89)
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.
