Deep learning and shape modelling for medical image reconstruction, segmentation and analysis

Daniel Rueckert
Imperial College

This talk will discuss deep learning approaches for the reconstruction, super-resolution and segmentation of Magnetic Resonance (MR) images. In particular, we will show how information about the shape of the anatomy can be incorporated as prior knowledge into these deep learning approaches. In addition, we show how shape and motion information can be used to develop interpretable deep learning approaches for diagnosis and prognosis.

Presentation (PDF File)

Back to Long Programs