Research map / Detection and segmentation / Open-set and few-shot detection
CoDet: Co-occurrence Guided Region-Word Alignment for Open-Vocabulary Object Detection
Chuofan Ma, Yi Jiang, Xin Wen, Zehuan Yuan, XIAOJUAN QI
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 text guidance to inter-region similarity (ablation) | CoDet · OV-COCO | AP50 novel: 26.6 -> 30.6 | Improved |
| prototype-based instead of heuristic co-occurring object discovery (ablation) | CoDet · OV-COCO | AP50 novel: 26.9 -> 30.6 | Improved |
| increase concept group size from 2 to 8 (ablation) | CoDet · OV-COCO | AP50 novel: 30.6 -> 29.1 | Hurt |
| increase concept group size from 2 to 8 (ablation) | CoDet · OV-LVIS | AP_novel (mask): 21.9 -> 22.7 | Improved |
| increase concept group size from 2 to 4 (ablation) | CoDet · OV-LVIS | AP_novel (mask): 21.9 -> 21.8 | No clear effect |
| switch backbone from ResNet50 to Swin-B | CoDet · OV-LVIS | AP_novel (mask): 23.4 -> 29.4 | Improved |
| switch backbone from ResNet50 to EVA02-L | CoDet · OV-LVIS | AP_novel (mask): 23.4 -> 37.0 | Improved |
| region-region similarity alignment vs region-word similarity self-training | CoDet · OV-COCO pseudo-labels | cover rate | Improved |
| compare CoDet to VLDet on OV-COCO | CoDet · OV-COCO | AP50 novel: 32.0 -> 30.6 | Hurt |
| cross-dataset transfer vs ViLD on COCO | CoDet · COCO transfer | AP: 36.6 -> 39.1 | Improved |
| cross-dataset transfer vs ViLD on Objects365 | CoDet · Objects365 transfer | AP: 11.8 -> 14.2 | Improved |
Closest papers by meaning
- Benefit From Seen: Enhancing Open-Vocabulary Object Detection by Bridging Visual and Textual Co-Occurrence Knowledge (ICCV 2025 · similarity 0.93)
- Exploring Region-Word Alignment in Built-in Detector for Open-Vocabulary Object Detection (CVPR 2024 · similarity 0.91)
- LLMs Meet VLMs: Boost Open Vocabulary Object Detection with Fine-grained Descriptors (ICLR 2024 · similarity 0.90)
- CLIM: Contrastive Language-Image Mosaic for Region Representation (AAAI 2024 · similarity 0.89)
- LLMDet: Learning Strong Open-Vocabulary Object Detectors under the Supervision of Large Language Models (CVPR 2025 · similarity 0.88)
- Cyclic Contrastive Knowledge Transfer for Open-Vocabulary Object Detection (ICLR 2025 · similarity 0.88)
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.
