Computer Vision News - August 2020

Sriprabha Ramanarayanan 27 Best of MIDL 2020 Sriprabha explains the computer vision part of thework: “ We use a deep learning network forMRI reconstructioncalled the Deep Cascaded Convolutional Neural Network , together with a Dynamic Weight Prediction network . These two networks function together and provide what is called a multiple acquisition context-based MRI reconstruction algorithm. This new network provides generalizability and flexibility in image reconstruction . ” The model has already shown promising results compared to existing methods and the plan is to develop it even further. They are currently working on a journal extension to improve the architecture to suit more clinical scenarios and extend the model to complex and parallel MRI datasets . “I’m very happy that the work has been acknowledged at MIDL 2020 and it has motivated me,” Sriprabha tells us. “We use a deep learning network for MRI reconstruction called the Deep Cascaded Convolutional Neural Network, together with a Dynamic Weight Prediction network”