SgVA-CLIP: Semantic-Guided Visual Adapting of Vision-Language Models for Few-Shot Image Classification | IEEE Journals & Magazine | IEEE Xplore