Research map / Detection and segmentation / Open-set and few-shot detection
OvarNet: Towards Open-Vocabulary Object Attribute Recognition
Keyan Chen, Xiaolong Jiang, Yao Hu, Xu Tang, Yan Gao, Jianqi Chen, Weidi Xie
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 parent-attribute words and learnable prompt vectors to the text prompt (ablation) | CLIP-Attr · VAW box-given | AP_novel: 53.78 -> 57.39 | Improved |
| Finetune with COCO-base and VAW-base annotations (ablation) | CLIP-Attr · VAW box-given | AP_novel: 46.15 -> 57.39 | Improved |
| Add image-caption training using COCO-Cap-sub (ablation) | CLIP-Attr · VAW box-given | AP_all: 66.92 -> 69.03 | Improved |
| Add the caption, noun-phrase, and largest-box attribute supervision signals (ablation) | CLIP-Attr Step-II · VAW box-given | AP_all: 66.92 -> 69.33 | Improved |
| Add top-K proposal attribute supervision to Step-II (ablation) | CLIP-Attr Step-II · COCO box-given | AP_all: 63.80 -> 65.17 | Improved |
| Distill attribute prediction probabilities with KL loss (ablation) | OvarNet · VAW box-given | AP_novel: 50.53 -> 56.43 | Improved |
| Use probability KL distillation instead of feature L2 distillation (ablation) | OvarNet · COCO box-given | AP_novel: 33.17 -> 54.10 | Improved |
| Initialize the visual backbone from CLIP-Attr rather than ImageNet (ablation) | OvarNet · VAW box-free | AP_novel: 50.27 -> 54.85 | Improved |
| Freeze rather than finetune the CLIP-Attr-initialized visual backbone (ablation) | OvarNet · VAW box-free | AP_novel: 54.85 -> 55.47 | No clear effect |
| Replace CNN blocks with Transformer attentional pooling (ablation) | OvarNet · VAW box-given | AP_novel: 53.35 -> 56.43 | Improved |
| Use 4 instead of 2 Transformer encoder layers for attentional pooling (ablation) | OvarNet · VAW box-given | AP_novel: 55.02 -> 56.43 | Improved |
| Vary the number of learnable prompt vectors (ablation) | CLIP-Attr · VAW box-given | AP_novel: 56.94 (3 prompts) -> 57.41 (60 prompts) | No clear effect |
Closest papers by meaning
- Open-Vocabulary Attribute Detection (CVPR 2023 · similarity 0.92)
- GUIDED: Granular Understanding via Identification, Detection, and Discrimination for Fine-Grained Open-Vocabulary Object Detection (NeurIPS 2025 · similarity 0.90)
- OW-VAP: Visual Attribute Parsing for Open World Object Detection (ICML 2025 · similarity 0.90)
- Compositional Caching for Training-free Open-vocabulary Attribute Detection (CVPR 2025 · similarity 0.89)
- The Devil is in the Fine-Grained Details: Evaluating Open-Vocabulary Object Detectors for Fine-Grained Understanding (CVPR 2024 · similarity 0.89)
- Open-Det: An Efficient Learning Framework for Open-Ended Detection (ICML 2025 · 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.
