3DConvCaps: 3DUnet with Convolutional Capsule Encoder for Medical Image Segmentation

TL;DR

This paper introduces 3DConvCaps, a novel 3D encoder-decoder network combining convolutional layers and capsule layers, which improves robustness and accuracy in medical image segmentation over traditional CNNs and previous capsule networks.

Contribution

The paper presents a new 3D capsule-based encoder-decoder architecture that effectively captures both local and global features for medical image segmentation.

Findings

Outperforms previous capsule networks and 3D-UNets on multiple datasets

Demonstrates improved robustness to pose and deformation

Shows effectiveness of combining convolutional and capsule layers

Abstract

Convolutional Neural Networks (CNNs) have achieved promising results in medical image segmentation. However, CNNs require lots of training data and are incapable of handling pose and deformation of objects. Furthermore, their pooling layers tend to discard important information such as positions as well as CNNs are sensitive to rotation and affine transformation. Capsule network is a recent new architecture that has achieved better robustness in part-whole representation learning by replacing pooling layers with dynamic routing and convolutional strides, which has shown potential results on popular tasks such as digit classification and object segmentation. In this paper, we propose a 3D encoder-decoder network with Convolutional Capsule Encoder (called 3DConvCaps) to learn lower-level features (short-range attention) with convolutional layers while modeling the higher-level features (long-range dependence) with capsule layers. Our experiments on multiple datasets including iSeg-2017, Hippocampus, and Cardiac demonstrate that our 3D 3DConvCaps network considerably outperforms previous capsule networks and 3D-UNets. We further conduct ablation studies of network efficiency and segmentation performance under various configurations of convolution layers and capsule layers at both contracting and expanding paths.