Reward-Guided Semantic Evolution for Test-time Adaptive Object Detection
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2026
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866911652224434176 |
|---|---|
| author | Zhou, Lihua Ye, Mao Zhu, Xiatian Li, Nianxin Ma, Changyi Li, Shuaifeng Qin, Yitong Liu, Hongbin Luo, Jiebo Lei, Zhen |
| author_facet | Zhou, Lihua Ye, Mao Zhu, Xiatian Li, Nianxin Ma, Changyi Li, Shuaifeng Qin, Yitong Liu, Hongbin Luo, Jiebo Lei, Zhen |
| contents | Open-vocabulary object detection with vision-language models (VLMs) such as Grounding DINO suffers from performance degradation under test-time distribution shifts, primarily due to semantic misalignment between text embeddings and shifted visual embeddings of region proposals. While recent test-time adaptive object detection methods for VLM-based either rely on costly backpropagation or bypass semantic misalignment via external memory, none directly and efficiently align text and vision in a training-free manner. To address this, we propose Reward-Guided Semantic Evolution (RGSE), a training-free framework that directly refines the text embeddings at test time. Inspired by evolutionary search, RGSE treats text embedding adaptation as a semantic search process: it perturbs text embeddings as candidate variants, evaluates them via cosine similarity with current and historical high-confidence visual proposals as a reward signal, and fuses them into a refined embedding through reward-weighted averaging. Without any backpropagation, RGSE achieves state-of-the-art performance across multiple detection benchmarks while adding minimal computational overhead. Our code will be open source upon publication. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_04531 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Reward-Guided Semantic Evolution for Test-time Adaptive Object Detection Zhou, Lihua Ye, Mao Zhu, Xiatian Li, Nianxin Ma, Changyi Li, Shuaifeng Qin, Yitong Liu, Hongbin Luo, Jiebo Lei, Zhen Computer Vision and Pattern Recognition Open-vocabulary object detection with vision-language models (VLMs) such as Grounding DINO suffers from performance degradation under test-time distribution shifts, primarily due to semantic misalignment between text embeddings and shifted visual embeddings of region proposals. While recent test-time adaptive object detection methods for VLM-based either rely on costly backpropagation or bypass semantic misalignment via external memory, none directly and efficiently align text and vision in a training-free manner. To address this, we propose Reward-Guided Semantic Evolution (RGSE), a training-free framework that directly refines the text embeddings at test time. Inspired by evolutionary search, RGSE treats text embedding adaptation as a semantic search process: it perturbs text embeddings as candidate variants, evaluates them via cosine similarity with current and historical high-confidence visual proposals as a reward signal, and fuses them into a refined embedding through reward-weighted averaging. Without any backpropagation, RGSE achieves state-of-the-art performance across multiple detection benchmarks while adding minimal computational overhead. Our code will be open source upon publication. |
| title | Reward-Guided Semantic Evolution for Test-time Adaptive Object Detection |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2605.04531 |