Few-Shot Image Classification and Segmentation as Visual Question Answering Using Vision-Language Models

no code implementations • 15 Mar 2024 • Tian Meng, Yang Tao, Ruilin Lyu, Wuliang Yin

By enabling a VLM to interact with off-the-shelf vision models as tools, the proposed method is capable of classifying and segmenting target objects using only image-level labels.

Few-Shot Image Classification In-Context Learning +3