Revisiting Integration of Image and Metadata for DICOM Series Classification: Cross-Attention and Dictionary Learning
Researchers have developed a new multimodal framework for classifying DICOM image series, integrating both image content and acquisition metadata. This approach uses a bi-directional cross-modal attention mechanism and a metadata encoder that handles missing or inconsistent data without imputation. The system is designed to manage variable series lengths and image dimensions, demonstrating superior performance over existing methods on both in-domain and out-of-domain evaluations. AI
IMPACT This new framework could improve the accuracy and efficiency of medical image analysis pipelines.