Research map / Detection and segmentation / Open-set and few-shot detection
Open Vocabulary Object Detection with Pseudo Bounding-Box Labels
Mingfei Gao, Chen Xing, Juan Carlos Niebles, Junnan Li, Ran Xu, Wenhao Liu, Caiming Xiong
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 Mask-RCNN proposals instead of Selective Search (ablation) | PB-OVD · COCO novel (fine-tuned) | novel AP: 28.5 -> 30.8 | Improved |
| use ALBEF instead of LXMERT for pseudo labels (ablation) | PB-OVD · COCO novel (fine-tuned) | novel AP: 27.0 -> 30.8 | Improved |
| use ALBEF instead of LXMERT for pseudo labels (ablation) | PB-OVD · LVIS (transfer) | AP50: 6.4 -> 8.0 | Improved |
| use 1.5k+ vocabulary instead of 65 COCO categories (ablation) | PB-OVD · COCO novel (fine-tuned) | novel AP: 29.7 -> 30.8 | Improved |
| use 1.5k+ vocabulary instead of 65 COCO categories (ablation) | PB-OVD · Objects365 (transfer) | AP50: +2.3% | Improved |
| use 1.5k+ vocabulary instead of 65 COCO categories (ablation) | PB-OVD · LVIS (transfer) | AP50: +2.7% | Improved |
| generate pseudo labels from COCO Cap, VG and SBU instead of COCO Cap only (ablation) | PB-OVD · COCO novel (fine-tuned) | novel AP: 29.1 -> 30.8 | Improved |
| use CLIP instead of Bert text encoder (ablation) | PB-OVD · COCO novel (fine-tuned) | novel AP: 28.8 -> 30.8 | Improved |
| Bert encoder PB-OVD vs Zareian et al. (Bert) | PB-OVD · COCO novel (fine-tuned) | novel AP: 22.8 -> 28.8 | Improved |
| fine-tune on COCO base categories after pseudo-label training (ablation) | PB-OVD · PASCAL VOC (transfer) | AP50: 44.4 -> 59.2 | Improved |
| fine-tune on COCO base categories after pseudo-label training (ablation) | PB-OVD · COCO novel (fine-tuned) | novel AP: 25.8 -> 30.8 | Improved |
| pre-train Zareian et al. VL model with more data (COCO Cap, VG, SBU) (ablation) | Zareian et al. · COCO novel (fine-tuned) | novel AP: -4.2% | Hurt |
Closest papers by meaning
- Generative Region-Language Pretraining for Open-Ended Object Detection (CVPR 2024 · similarity 0.90)
- Taming Self-Training for Open-Vocabulary Object Detection (CVPR 2024 · similarity 0.90)
- Language-conditioned Detection Transformer (CVPR 2024 · similarity 0.89)
- ProxyDet: Synthesizing Proxy Novel Classes via Classwise Mixup for Open-Vocabulary Object Detection (AAAI 2024 · similarity 0.89)
- Exploring Region-Word Alignment in Built-in Detector for Open-Vocabulary Object Detection (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.
