The Translational Imaging Group has recently published their work on deep learning-based interactive image segmentation in IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI). The segmentation method, named as DeepIGeoS, comebines deep convolutional neural networks with user interactions for 2D and 3D image segmentation.

This method is fast and efficient, and can achieve accurate segmentation with only few user interactions, which saves a lot of user time compared with traditional interactive segmentation methods. The authors have applied DeepIGeoS to 2D placenta segmentation from fetal MRI and 3D brain tumor segmentation from FLAIR images.


Guotai Wang, Maria A Zuluaga, Wenqi Li, Rosalind Pratt, Premal A Patel, Michael Aertsen, Tom Doel, Anna L David, Jan Deprest, Sebastien Ourselin, Tom Vercauteren. DeepIGeoS: A Deep Interactive Geodesic Framework for Medical Image Segmentation. IEEE Transactions on Pattern Analysis and Machine Intelligence, (In press) 2018. doi: 10.1109/TPAMI.2018.2840695.


A video showing DeepIGeoS for 2D segmentation: video1.mp4

A video showing DeepIGeoS for 3D segmentation: video2.mp4