Efficient Folded Attention for 3D Medical Image Reconstruction and Segmentation
This work addresses efficiency issues in 3D medical image analysis for researchers and practitioners, though it is incremental as it builds on existing attention methods.
The paper tackled the computational challenge of applying attention mechanisms to large 3D medical images by proposing a folded attention approach, which reduced computational complexity and GPU memory consumption while moderately improving accuracy on tasks like quantitative susceptibility mapping and multiple sclerosis lesion segmentation.
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.