Research map / Detection and segmentation / Open-set and few-shot detection
HOPS: Hierarchical Open-vocabulary Part Segmentation with Attention-Aware Filtering and Affinity-Guided Enhancement
Xinlong Li, Di Lin, Shaoyiyi Gao, Yaxuan Liu, Jixian He, Jiaxin Li, Ruonan Liu, Qing Guo, Kairui Yang, Wei Feng
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 AFM to frozen CLIP object segmentation (ablation) | CLIP-freeze · PASCAL VOC (all classes) | mOSI: 42.33 -> 31.45 | Improved |
| add AFM to supervised CLIP-train (ablation) | CLIP-train · PASCAL VOC (all classes) | h-IoU: 62.45 -> 66.53 | Improved |
| supervised training of CLIP on unseen classes (ablation) | CLIP-train vs CLIP-freeze · PASCAL VOC (unseen) | mIoU: 63.37 -> 56.94 | Hurt |
| two-stage (object then part) instead of one-stage part segmentation (ablation) | frozen CLIP · Pascal-Part-116 | mIoU: 18.06 -> 23.42 | Improved |
| add AFM to two-stage part segmentation (ablation) | two-stage frozen CLIP · Pascal-Part-116 | mIoU: 23.42 -> 27.52 | Improved |
| add AEM on top of AFM (ablation) | two-stage frozen CLIP · Pascal-Part-116 | mIoU: 27.52 -> 30.28 | Improved |
| add semantic-structural attention to AFM (ablation) | HOPS · PartImageNet | h-IoU: 56.37 -> 57.02 | Improved |
| add reliable association mask to AFM (ablation) | HOPS · PartImageNet | h-IoU: 57.02 -> 57.86 | Improved |
| use full AFM (all three components) (ablation) | HOPS · PartImageNet | mOSI: 24.92 -> 19.08 | Improved |
| add semantic-structural affinity matrix to AEM (ablation) | HOPS · PartImageNet | mUSI: 34.24 -> 33.18 | Improved |
| add bounding box mask constraint to AEM propagation (ablation) | HOPS · PartImageNet | mUSI: 33.18 -> 31.76 | Improved |
| adaptive iteration without bounding box (ablation) | HOPS AEM · PartImageNet | mUSI: 33.18 -> 32.41 | Improved |
Closest papers by meaning
- Fine-Grained Image-Text Correspondence with Cost Aggregation for Open-Vocabulary Part Segmentation (CVPR 2025 · similarity 0.89)
- LangHOPS: Language Grounded Hierarchical Open-Vocabulary Part Segmentation (NeurIPS 2025 · similarity 0.89)
- Understanding Multi-Granularity for Open-Vocabulary Part Segmentation (NeurIPS 2024 · similarity 0.88)
- Emergent Open-Vocabulary Semantic Segmentation from Off-the-shelf Vision-Language Models (CVPR 2024 · similarity 0.88)
- Training-free Open-Vocabulary Semantic Segmentation via Diverse Prototype Construction and Sub-region Matching (AAAI 2025 · similarity 0.87)
- OpenWorldSAM: Extending SAM2 for Universal Image Segmentation with Language Prompts (NeurIPS 2025 · similarity 0.87)
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.
