Deep supervision network with contrastive learning for zero-shot sketch-based image retrieval

Published: 01 Jan 2024, Last Modified: 10 Apr 2025Appl. Soft Comput. 2024EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•The proposed method reduces the difficulty of optimizing the intermediate layers.•It bridges the domain and semantic gaps by aligning sketch and image features.•It overcomes the limitation of batch size in metric learning.
Loading