Revisiting Few-Shot Object Detection with Vision-Language Models

arXiv:2312.14494v2 Announce Type: replace
Abstract: Few-shot object detection (FSOD) benchmarks have advanced techniques for detecting new categories with limited annotations. Existing benchmarks repurpose well-established datasets like COCO by partitioning categories into base and novel classes for pre-training and fine-tuning respectively. However, these benchmarks do not reflect how FSOD is deployed in practice. Rather than only pre-training on a small number of base categories, we argue that it is more practical to fine-tune a foundation model (e.g., a vision-language model (VLM) …

arxiv cs.cv detection few-shot language language models object type vision vision-language vision-language models