ABRA: Teleporting Fine-Tuned Knowledge Across Domains for Open-Vocabulary Object Detection
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914390126624768 |
|---|---|
| author | Bernardi, Mattia Cappellino, Chiara Mosconi, Matteo Sangineto, Enver Porrello, Angelo Calderara, Simone |
| author_facet | Bernardi, Mattia Cappellino, Chiara Mosconi, Matteo Sangineto, Enver Porrello, Angelo Calderara, Simone |
| contents | Although recent Open-Vocabulary Object Detection architectures, such as Grounding DINO, demonstrate strong zero-shot capabilities, their performance degrades significantly under domain shifts. Moreover, many domains of practical interest, such as nighttime or foggy scenes, lack large annotated datasets, preventing direct fine-tuning. In this paper, we introduce Aligned Basis Relocation for Adaptation(ABRA), a method that transfers class-specific detection knowledge from a labeled source domain to a target domain where no training images containing these classes are accessible. ABRA formulates this adaptation as a geometric transport problem in the weight space of a pretrained detector, aligning source and target domain experts to transport class-specific knowledge. Extensive experiments across challenging domain shifts demonstrate that ABRA successfully teleports class-level specialization under multiple adverse conditions. Our code will be made public upon acceptance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_12409 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | ABRA: Teleporting Fine-Tuned Knowledge Across Domains for Open-Vocabulary Object Detection Bernardi, Mattia Cappellino, Chiara Mosconi, Matteo Sangineto, Enver Porrello, Angelo Calderara, Simone Computer Vision and Pattern Recognition Although recent Open-Vocabulary Object Detection architectures, such as Grounding DINO, demonstrate strong zero-shot capabilities, their performance degrades significantly under domain shifts. Moreover, many domains of practical interest, such as nighttime or foggy scenes, lack large annotated datasets, preventing direct fine-tuning. In this paper, we introduce Aligned Basis Relocation for Adaptation(ABRA), a method that transfers class-specific detection knowledge from a labeled source domain to a target domain where no training images containing these classes are accessible. ABRA formulates this adaptation as a geometric transport problem in the weight space of a pretrained detector, aligning source and target domain experts to transport class-specific knowledge. Extensive experiments across challenging domain shifts demonstrate that ABRA successfully teleports class-level specialization under multiple adverse conditions. Our code will be made public upon acceptance. |
| title | ABRA: Teleporting Fine-Tuned Knowledge Across Domains for Open-Vocabulary Object Detection |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2603.12409 |