Efficient Folded Attention for 3D Medical Image Reconstruction and Segmentation

Abstract

Recently, 3D medical image reconstruction (MIR) and segmentation (MIS) based on deep neural networks have been developed with promising results, and attention mechanism has been further designed to capture global contextual information for performance enhancement. However, the large size of 3D volume images poses a great computational challenge to traditional attention methods. In this paper, we propose a folded attention (FA) approach to improve the computational efficiency of traditional attention methods on 3D medical images. The main idea is that we apply tensor folding and unfolding operations with four permutations to build four small sub-affinity matrices to approximate the original affinity matrix. Through four consecutive sub-attention modules of FA, each element in the feature tensor can aggregate spatial-channel information from all other elements. Compared to traditional attention methods, with moderate improvement of accuracy, FA can substantially reduce the computational complexity and GPU memory consumption. We demonstrate the superiority of our method on two challenging tasks for 3D MIR and MIS, which are quantitative susceptibility mapping and multiple sclerosis lesion segmentation.

arXiv comment9 pages, 7 figures

References in corpus (11)

• Adam: A Method for Stochastic Optimization
• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
• Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
• Neural Machine Translation by Jointly Learning to Align and Translate
• Attention U-Net: Learning Where to Look for the Pancreas
• Squeeze-and-Excitation Networks
• CCNet: Criss-Cross Attention for Semantic Segmentation
• Geometric Loss for Deep Multiple Sclerosis lesion Segmentation
• Fidelity Imposed Network Edit (FINE) for Solving Ill-Posed Image Reconstruction
• QSMGAN: Improved Quantitative Susceptibility Mapping using 3D Generative Adversarial Networks with Increased Receptive Field

Cited by in corpus (3)

• Spatially Covariant Lesion Segmentation
• mcLARO: Multi-Contrast Learned Acquisition and Reconstruction Optimization for simultaneous quantitative multi-parametric mapping
• DeDA: Deep Directed Accumulator