Weakly-Supervised Convolutional Neural Networks for Multimodal Image Registration

Multimodal image registration aims to spatially align medical images produced from different imaging modalities. Among many other medical imaging applications, this is useful in minimally- or none-invasive image-guided procedures, in which a common strategy is to fuse the detailed diagnostic information from quality pre-procedural images with intra-procedural imaging that is typically restricted by the interventional requirements, such as portability, accessibility, temporal resolution, limited field of view and controlled dosage for contrast agent or radiation.

Source: Medical Image Analysis - July 4, 2018 Category: Radiology Authors: Yipeng Hu, Marc Modat, Eli Gibson, Wenqi Li, Nooshin Ghavami, Ester Bonmati, Guotai Wang, Steven Bandula, Caroline M. Moore, Mark Emberton, S ébastien Ourselin, J. Alison Noble, Dean C. Barratt, Tom Vercauteren Source Type: research