CVPR Daily - Wednesday

DAILY Wednesday Workshop 16 A number of exciting medical industry talks were presented. The workshop began with a great talk by Ayelet Akselrod-Ballin and Raouf Muhamedrahimov , from Zebra Medical Vision Ltd. They described methods for classification of Intravenous Contrast enhancement phases in CT images, which jointly learn categorical label representations and can implicitly encode natural interclass relations. They described the process of generating a product from an algorithm, emphasizing the unique aspects of research in the medical imaging industry. There were a number of fascinating presentations from companies focused on machine learning for computational pathology in the context of cancer care. Chris Katan , PaigeAI, gave a very interesting talk showcasing the results with PaigeAI’s deep learning based systems for multiple tissue to assist pathologists in finding cancer and describing the challenges in generalizing deep learning models across hospitals. Martin Stumpe , Tempus, described opportunities, challenges and proposing solutions for developing machine learning for precision medicine for cancer, with the goal of making cancer diagnosis and prognosis more accurate and consistent. Aïcha BenTaieb , also from Tempus, showed some work on computational tools to detect and diagnose cancers from histopathology tissue slides, specifically to identify histology-derived biomarkers to predict patients' response to treatment & risks of disease recurrence. Dakai Jim , PAII, described some recent advances in parsing and segmenting structures for head-and neck cancer radiation therapy planning.