Deep global-ranking hashing via average precision approximation for large-scale image retrieval

Key Points

• Enhanced average precision achieved with the proposed deep hashing algorithm, targeting large-scale image datasets.
• Results show significant improvements in retrieval performance, outperforming traditional methods by over 30% in specific tests.
• Analytical framework was developed to approximate average precision in image retrieval tasks, validating its effectiveness.
• Implications suggest potential applications in various domains, like e-commerce and social media, to improve user experience.