About this document
Elevating Zero-Shot Sketch-Based Image Retrieval Through Multimodal Prompt Learning by tranductdn is a document available to read on EtoBox.
The document presents SpLIP, a novel multimodal prompt learning approach for zero-shot sketch-based image retrieval (SBIR) that leverages the CLIP model. It introduces a bi-directional prompt-sharing strategy to enhance the synergy between visual and textual encoders, significantly improving the retrieval performance across various SBIR scenarios. The proposed method includes innovative techniques such as adaptive margin generation for triplet loss and a conditional cross-modal jigsaw task to refine sketch-
- Author
- tranductdn
- Language
- EN