File Download
There are no files associated with this item.
Supplementary
-
Citations:
- Appears in Collections:
Conference Paper: CoDet: Co-Occurrence Guided Region-Word Alignment for Open-Vocabulary Object Detection
Title | CoDet: Co-Occurrence Guided Region-Word Alignment for Open-Vocabulary Object Detection |
---|---|
Authors | |
Issue Date | 10-Dec-2023 |
Abstract | Deriving reliable region-word alignment from image-text pairs is critical to learn object-level vision-language representations for open-vocabulary object detection. Existing methods typically rely on pre-trained or self-trained vision-language models for alignment, which are prone to limitations in localization accuracy or generalization capabilities. In this paper, we propose CoDet, a novel approach that overcomes the reliance on pre-aligned vision-language space by reformulating region-word alignment as a co-occurring object discovery problem. Intuitively, by grouping images that mention a shared concept in their captions, objects corresponding to the shared concept shall exhibit high co-occurrence among the group. CoDet then leverages visual similarities to discover the co-occurring objects and align them with the shared concept. Extensive experiments demonstrate that CoDet has superior performances and compelling scalability in open-vocabulary detection, e.g., by scaling up the visual backbone, CoDet achieves 37.0 and 44.7 on OV-LVIS, surpassing the previous SoTA by 4.2 and 9.8 . Code is available at https://github.com/CVMI-Lab/CoDet. |
Persistent Identifier | http://hdl.handle.net/10722/340352 |
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Ma, Chuofan | - |
dc.contributor.author | Jiang, Yi | - |
dc.contributor.author | Wen, Xin | - |
dc.contributor.author | Yuan, Zehuan | - |
dc.contributor.author | Qi, Xiaojuan | - |
dc.date.accessioned | 2024-03-11T10:43:31Z | - |
dc.date.available | 2024-03-11T10:43:31Z | - |
dc.date.issued | 2023-12-10 | - |
dc.identifier.uri | http://hdl.handle.net/10722/340352 | - |
dc.description.abstract | <p>Deriving reliable region-word alignment from image-text pairs is critical to learn object-level vision-language representations for open-vocabulary object detection. Existing methods typically rely on pre-trained or self-trained vision-language models for alignment, which are prone to limitations in localization accuracy or generalization capabilities. In this paper, we propose CoDet, a novel approach that overcomes the reliance on pre-aligned vision-language space by reformulating region-word alignment as a co-occurring object discovery problem. Intuitively, by grouping images that mention a shared concept in their captions, objects corresponding to the shared concept shall exhibit high co-occurrence among the group. CoDet then leverages visual similarities to discover the co-occurring objects and align them with the shared concept. Extensive experiments demonstrate that CoDet has superior performances and compelling scalability in open-vocabulary detection, e.g., by scaling up the visual backbone, CoDet achieves 37.0 and 44.7 on OV-LVIS, surpassing the previous SoTA by 4.2 and 9.8 . Code is available at https://github.com/CVMI-Lab/CoDet.</p> | - |
dc.language | eng | - |
dc.relation.ispartof | Neural Information Processing Systems 2023 (10/12/2023-16/12/2023, , , New Orleans) | - |
dc.title | CoDet: Co-Occurrence Guided Region-Word Alignment for Open-Vocabulary Object Detection | - |
dc.type | Conference_Paper | - |