Research map / Detection and segmentation / Open-set and few-shot detection
Diffusion Models for Open-Vocabulary Segmentation
Laurynas Karazija, Iro Laina, Andrea Vedaldi, Christian Rupprecht
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 background prototypes (ablation) | OVDiff (SD features) · PASCAL VOC | mIoU: 53.2 -> 64.4 | Improved |
| add category pre-filtering (ablation) | OVDiff (SD features) · PASCAL VOC | mIoU: 54.4 -> 64.4 | Improved |
| add CutLER for mask extraction (ablation) | OVDiff (SD features) · PASCAL VOC | mIoU: 60.4 -> 64.4 | Improved |
| add sliding window inference (ablation) | OVDiff · PASCAL VOC | mIoU: 62.2 -> 64.4 | Improved |
| add instance and part prototypes to average prototype (ablation) | OVDiff · PASCAL VOC | mIoU: 62.5 -> 64.4 | Improved |
| ensemble SD, CLIP and DINO features instead of SD alone (ablation) | OVDiff · PASCAL VOC | mIoU: 64.4 -> 66.4 | Improved |
| use CLIP keys instead of CLIP patch tokens (ablation) | OVDiff · PASCAL VOC | mIoU: 51.4 -> 61.8 | Improved |
| add category filter to prior methods TCL and OVSegmentor (ablation) | TCL, OVSegmentor · PASCAL VOC | mIoU: TCL 51.2 -> 51.8; OVSegmentor 53.8 -> 53.7 | No clear effect |
| add CutLER majority voting to prior methods (ablation) | TCL, OVSegmentor · PASCAL VOC | mIoU: TCL 51.2 -> 43.1; OVSegmentor 53.8 -> 38.7 | Hurt |
| replace CutLER with supervised SAM or Grounded SAM (ablation) | OVDiff · PASCAL VOC | mIoU: 66.3 -> 68.5 | Improved |
| use newer Stable Diffusion generator (SD 2.0 vs 1.5) (ablation) | OVDiff · PASCAL VOC | mIoU: 66.4 -> 67.7 | Improved |
| vary number of part prototype clusters K from 8 to 64 (ablation) | OVDiff · PASCAL VOC | mIoU: 63.8 -> 64.4 (K=32) | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- GroupViT — GroupViT: Semantic Segmentation Emerges From Text Supervision (CVPR 2022) · 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
- Emergent Open-Vocabulary Semantic Segmentation from Off-the-shelf Vision-Language Models (CVPR 2024 · similarity 0.91)
- Training-free Open-Vocabulary Semantic Segmentation via Diverse Prototype Construction and Sub-region Matching (AAAI 2025 · similarity 0.91)
- Transferable and Principled Efficiency for Open-Vocabulary Segmentation (CVPR 2024 · similarity 0.91)
- Training-Free Open-Vocabulary Segmentation with Offline Diffusion-Augmented Prototype Generation (CVPR 2024 · similarity 0.91)
- Open-Vocabulary Attention Maps with Token Optimization for Semantic Segmentation in Diffusion Models (CVPR 2024 · similarity 0.90)
- USE: Universal Segment Embeddings for Open-Vocabulary Image Segmentation (CVPR 2024 · 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.
