Agenda

Séminaire ICE « Computational Imaging for Medical Diagnosis »

Abstract

The significant progress in terms of computing power, automated data processing, but also sensor technologies obtained over the last 20 years have undeniably participated to the development of innovative imaging systems. Among these, computational imaging plays a singular role because it can exceed the capabilities of traditional imaging in terms of optical resolution but also in terms of information acquired on the tissues that is analyzed. This type of imaging relies largely on the resolution of inverse problems (light-matter interaction, image formation). Their operation and their interest will be presented in relation to various challenges posed to the digital pathology sector. The contribution of deep-learning for the resolution of inverse problems (physics-informed deep-learning) will also be discussed.

Bio

Yaneck Gottesman is professor at telecom SudParis since 2002. His research activities were initially focused on photonics integrated circuits, optical characterization, nano-photonics, and instrumental optics. Since 2015, he has been interested in the imaging of living organisms. For this purpose, he developed a computational imaging platform dedicated to the analysis of biological tissues. This open platform brings together different instruments conceived by a TSP/SAMOVAR interdisciplinary team of opticians and machine learning/data processing researchers. He is at the origin of a highly secure biometric sensor and is currently contributing to the development of unconventional imaging approaches in optical microscopy.