【論文要約:自動運転関連】Revisiting Few-Shot Object Detection with Vision-Language Models