FACTOR: Counterfactual Training-Free Test-Time Adaptation for Open-Vocabulary Object Detection
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866914528719011840 |
|---|---|
| author | Zhao, Kaixiang Ye, Mao Zhou, Lihua Wang, Hu Ji, Luping Tang, Song Zhu, Xiatian |
| author_facet | Zhao, Kaixiang Ye, Mao Zhou, Lihua Wang, Hu Ji, Luping Tang, Song Zhu, Xiatian |
| contents | Open-vocabulary object detection often fails under distribution shifts, as it can be misled by spurious correlations between non-causal visual attributes (e.g., brightness, texture) and object categories. Existing test-time adaptation (TTA) methods either depend on costly online optimization or perform global calibration, overlooking the attribute-specific nature of these failures. To address this, we propose FACTOR (counterFACtual training-free Test-time adaptation for Open-vocabulaRy object detection), a lightweight framework grounded in counterfactual reasoning. By perturbing test images along non-causal attributes and comparing region-level predictions between original and counterfactual views, FACTOR quantifies attribute sensitivity, semantic relevance, and prediction variation to selectively suppress attribute-dependent predictions-without parameter updates. Experiments on PASCAL-C, COCO-C, and FoggyCityscapes show that FACTOR consistently outperforms prior TTA methods, demonstrating that explicit counterfactual reasoning effectively improves robustness under distribution shifts. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_03294 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | FACTOR: Counterfactual Training-Free Test-Time Adaptation for Open-Vocabulary Object Detection Zhao, Kaixiang Ye, Mao Zhou, Lihua Wang, Hu Ji, Luping Tang, Song Zhu, Xiatian Computer Vision and Pattern Recognition Open-vocabulary object detection often fails under distribution shifts, as it can be misled by spurious correlations between non-causal visual attributes (e.g., brightness, texture) and object categories. Existing test-time adaptation (TTA) methods either depend on costly online optimization or perform global calibration, overlooking the attribute-specific nature of these failures. To address this, we propose FACTOR (counterFACtual training-free Test-time adaptation for Open-vocabulaRy object detection), a lightweight framework grounded in counterfactual reasoning. By perturbing test images along non-causal attributes and comparing region-level predictions between original and counterfactual views, FACTOR quantifies attribute sensitivity, semantic relevance, and prediction variation to selectively suppress attribute-dependent predictions-without parameter updates. Experiments on PASCAL-C, COCO-C, and FoggyCityscapes show that FACTOR consistently outperforms prior TTA methods, demonstrating that explicit counterfactual reasoning effectively improves robustness under distribution shifts. |
| title | FACTOR: Counterfactual Training-Free Test-Time Adaptation for Open-Vocabulary Object Detection |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2605.03294 |