AutoConcept: Training-Free Concept-Guided Reranking for Metadata-Available Composed Image Retrieval
This paper introduces AutoConcept, a training-free reranker that converts concept evidence into an interpretable memory that positions AutoConcept as an interpretable concept-memory reranker for product-style CIR galleries with available metadata.
Tianyi Wang, Tian-Jiao Wu
· 0 citations