Cross-modal hashing retrieval approaches maps heterogeneous multi-modal data into a common hamming space to achieve efficient and flexible retrieval performance. However, existing cross-modal methods mainly exploit feature-level similarity between multi-modal data, the label-level similarity and relative ranking relationship between adjacent instances have been ignored. To address these problems, we propose a novel Deep Rank Cross-modal Hashing(DRCH) method that fully explores the intra-modal semantic similarity relationship.

We propose a new method called DCH-SCR for cross-modal retrieval, which addresses issues with existing methods that only consider feature-level similarity and ignore label-level fine-grained similarity. Our method preserves semantic similarity by combining label-level and feature-level information, narrows the gap between modalities with a ranking alignment loss function, and optimizes hash codes based on a common semantic space. We use the gradient and Normalized Discounted Cumulative Gain to achieve varying optimization strengths for data pairs with different similarities. Experiments on three image-text retrieval datasets show that DCH-SCR outperforms state-of-the-art methods.