Multi-omics and Imaging Cross-modal Feature Fusion Research for Precision Diagnosis and Treatment

This direction focuses on cross-modal feature fusion of multi-omics molecular data and medical imaging data as well as its applications in precision diagnosis and treatment. Aiming at the key challenges of strong data heterogeneity, ambiguous correlation mechanisms and insufficient fusion modeling accuracy between molecular omics and imaging phenotypic data, adaptive cross-modal feature fusion algorithms are developed. This research deeply explores the coupling relationship between microscopic molecular features and macroscopic imaging phenotypes, constructs cross-scale and high-generalization models for precise disease typing, early warning and prognostic prediction, and realizes comprehensive and accurate judgment from molecular mechanisms to clinical phenotypes, promoting the technological upgrading of precision medicine.