Undergraduate Thesis/Research — Multimodal integration gap in deep learning for intracranial aneurysm detection
DiagnosisLed research activities aimed at understanding and addressing how deep learning models integrate information from multiple imaging modalities for intracranial aneurysm detection. The thesis/work focused on characterizing the multimodal integration gap and how it impacts detection accuracy. This constitutes AI training/research work that relates to assembling, interpreting, and evaluating medical imaging model inputs. • Conducted systematic review work on deep learning multimodal integration for aneurysm detection • Assessed limitations of current models across imaging modalities • Synthesized findings to inform model development/evaluation considerations • Used literature-based evidence to guide how multimodal data should be handled in detection tasks