Thumbnails:
List:
Year:
Category:
Session:
Poster:
Getting poster data...
Piyush Kumar Singh, Theodoros Soldatos (SRH Heidelberg, Germany.)
While several biomedical imaging databases (DBs) exist, only few allow for efficient search, browsing and download options. Reasons for this include the vast amount of data that must be stored, indexed, analyzed and visualized every time. When it comes to querying, it is rather text-based searches that are more frequently allowed, rather than image-based. For this reason, we wondered whether we could help make such searches more focused. In this work, we demonstrate that this is feasible with our medical image and text organizer (MITO) approach. In specific, we show results from a proof of concept (PoC) work that combines (a) deep learning to filter the candidate biomedical imaging modality search space and (b) NLP topic modelling to summarize relevant literature. Our PoC shows that MITO can enable such an efficient biomedical image identifying and relevant literature extraction.