Computer Vision News - November 2016

Islem Rekik presented the work of Khosro Bahrami titled “convolutional neural network for reconstruction of 7T-like images from 3T MRI using appearance and anatomical features”. The 7T MRI scanners has higher resolution and contrast than the conventional 3T MRI and can potentially help in more accurate diagnosis of various brain diseases. Currently, 7T MRI scanners are less accessible compared to the 3T MRI scanners. They proposed a deep architecture for Convolutional Neural Network (CNN), which uses the appearance (intensity) and anatomical (labels of brain tissues) features as input to non- linearly map 3T MRI to 7T MRI. They showed that both visual and numerical results outperform different comparison methods. CVPR Daily: Thursday Workshop - DLMIA 57 by Avi Ben Cohen Convolutional Neural Network for Reconstruction of 7T-like Images from 3T MRI Using Appearance and Anatomical Features One particularity of MICCAI 2016 was that the number of participants to the workshops was even higher than the number of the attendees at the regular conference. This can be considered a proof of the quality of the workshop proposed. Our readers will not be surprised to learn that the most successful workshop at MICCAI was the 2nd Workshop on Deep Learning in Medical Image Analysis. Avi Ben Cohen was there and he tells us about one of the presentations at DLMIA 2016. BEST OF MICCAI