Computer Vision News - May 2018

ISBI DAILY Saturday 31 Min Tang In terms of algorithmic techniques, the detection module uses Region Proposal Networks and the segmentation module uses a 3D U-Net. Min tells us about the next steps of her work: “ This algorithm now can automatically segment the 3D volume. If you want to make it more accurate, instead of using the detection part you can use a human input. You just select the region of your interest and input to the 3D U-Net. That would be much more accurate than my method. It’s a human interactive method. ” “ Rather than segmenting an original image directly, it uses a detection module first to detect the region of interest. ” BEST OF