Joint asymmetric discrete hashing for cross-modal retrieval

Key Points

• Achieves improved similarity detection across different data types, enhancing retrieval quality.
• Experiments demonstrate a notable 30% increase in retrieval accuracy compared to conventional methods.
• Analysis involved methods using asymmetric embeddings to optimize cross-modal relationships between datasets.
• Significance may allow for better data management in applications requiring fast cross-modal searches.