Hierarchical 3D Feature Learning for Pancreas Segmentation
This work addresses the problem of accurate pancreas segmentation in medical imaging for healthcare applications, representing an incremental improvement with a novel method for a known bottleneck.
The authors tackled automated pancreas segmentation from MRI and CT scans using a novel 3D fully convolutional deep network, achieving an average Dice score of 88% on CT data and 77% on MRI data, outperforming existing methods on CT.
We propose a novel 3D fully convolutional deep network for automated pancreas segmentation from both MRI and CT scans. More specifically, the proposed model consists of a 3D encoder that learns to extract volume features at different scales; features taken at different points of the encoder hierarchy are then sent to multiple 3D decoders that individually predict intermediate segmentation maps. Finally, all segmentation maps are combined to obtain a unique detailed segmentation mask. We test our model on both CT and MRI imaging data: the publicly available NIH Pancreas-CT dataset (consisting of 82 contrast-enhanced CTs) and a private MRI dataset (consisting of 40 MRI scans). Experimental results show that our model outperforms existing methods on CT pancreas segmentation, obtaining an average Dice score of about 88%, and yields promising segmentation performance on a very challenging MRI data set (average Dice score is about 77%). Additional control experiments demonstrate that the achieved performance is due to the combination of our 3D fully-convolutional deep network and the hierarchical representation decoding, thus substantiating our architectural design.