Pancreas segmentation with probabilistic map guided bi-directional recurrent UNet
Pancreas segmentation in medical imaging data is of great significance for clinical pancreas diagnostics and treatment. However, the large population variations in the pancreas shape and volume cause enormous segmentation difficulties, even for state-of-the-art algorithms utilizing fully-convolutional neural networks (FCNs). Specifically, pancreas segmentation suffers from the loss of spatial information in 2D methods, and the high computational cost of 3D methods. To alleviate these problems, we propose a probabilistic-map-guided bi-directional recurrent UNet (PBR-UNet) architecture, which fuses intra-slice information and inter-slice probabilistic maps into a local 3D hybrid regularization scheme, which is followed by bi-directional recurrent network optimization. The PBR-UNet method consists of an initial estimation module for efficiently extracting pixel-level probabilistic maps and a primary segmentation module for propagating hybrid information through a 2.5D U-Net architecture. Specifically, local 3D information is inferred by combining an input image with the probabilistic maps of the adjacent slices into multichannel hybrid data, and then hierarchically aggregating the hybrid information of the entire segmentation network. Besides, a bi-directional recurrent optimization mechanism is developed to update the hybrid information in both the forward and the backward directions. This allows the proposed network to make full and optimal use of the local context information. Quantitative and qualitative evaluation was performed on the NIH Pancreas-CT dataset, and our proposed PBR-UNet method achieved better segmentation results with less computational cost compared to other state-of-the-art methods.