Recently, Prof. Zhang Huaiwen’s research team from the College of Computer Science ( College of Software) , College of Artificial Intelligence at IMU made new progress in the field of cross-modal retrieval. Their paper, titled “Active Supervised Cross-Modal Retrieval,” was published in IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). TPAMI is widely recognized as a top-tier international journal in the field of artificial intelligence, recommended as an A-class journal by the China Computer Federation (CCF), with an impressive impact factor of 20.8. Prof. Zhang Huaiwen and Researcher Yang Yang are co-first authors of the paper, with IMU listed as the primary contributing institution.

Focusing on the task of cross-modal retrieval, the paper addresses the high annotation costs associated with massive multimodal datasets in practical applications. The research proposes an active supervised cross-modal retrieval method, which identifies the most informative multimodal samples to enable unbiased sample selection. This approach achieves high-level cross-modal retrieval performance while significantly reducing annotation overhead.