Exploration of Different Attention Mechanisms on Medical Image Segmentation

Nowadays, medical image segmentation plays an important role in computer-aided medical diagnosis. To realize effective segmentation, Attention Mechanism (AM) is widely adopted. It can be trained to automatically highlight salient features and integrated into convolution neural networks conveniently. However, many researchers choose the attention mechanism without sufficient theoretical interpretability. They ignore the differences and dominant characteristics between various datasets, which causes the failure to select the most appropriate one. In this paper, we explore the implementation and discrimination of four specific attention mechanisms. To evaluate their performances, we incorporate these mechanisms within the U-Net and make a comparison on three medical image datasets. The experimental results show that all these attention mechanisms can improve the value of Mean IoU. More significantly, we find the best AM for each type of dataset and analyze the reasons for different performances from underlying mathematical principles.

[1]  Ronan Collobert,et al.  Recurrent Convolutional Neural Networks for Scene Labeling , 2014, ICML.

[2]  Surabhi Bhargava,et al.  A Dataset and a Technique for Generalized Nuclear Segmentation for Computational Pathology , 2017, IEEE Transactions on Medical Imaging.

[3]  Tat-Seng Chua,et al.  SCA-CNN: Spatial and Channel-Wise Attention in Convolutional Networks for Image Captioning , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[4]  Yi Yang,et al.  Attention to Scale: Scale-Aware Semantic Image Segmentation , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[5]  Jie Li,et al.  Channel-Wise and Spatial Feature Modulation Network for Single Image Super-Resolution , 2018, IEEE Transactions on Circuits and Systems for Video Technology.

[6]  Fernando Vilariño,et al.  WM-DOVA maps for accurate polyp highlighting in colonoscopy: Validation vs. saliency maps from physicians , 2015, Comput. Medical Imaging Graph..

[7]  Alex Graves,et al.  Recurrent Models of Visual Attention , 2014, NIPS.

[8]  Iasonas Kokkinos,et al.  DeepLab: Semantic Image Segmentation with Deep Convolutional Nets, Atrous Convolution, and Fully Connected CRFs , 2016, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[9]  Hong Zhang,et al.  Multi-scale Patch Aggregation (MPA) for Simultaneous Detection and Segmentation , 2016, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[10]  Thomas Brox,et al.  U-Net: Convolutional Networks for Biomedical Image Segmentation , 2015, MICCAI.

[11]  Honglak Lee,et al.  A unified framework for automatic wound segmentation and analysis with deep convolutional neural networks , 2015, 2015 37th Annual International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC).

[12]  Guosheng Lin,et al.  Efficient Piecewise Training of Deep Structured Models for Semantic Segmentation , 2015, 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[13]  Hai Su,et al.  Region segmentation in histopathological breast cancer images using deep convolutional neural network , 2015, 2015 IEEE 12th International Symposium on Biomedical Imaging (ISBI).

[14]  Jun Fu,et al.  Dual Attention Network for Scene Segmentation , 2018, 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).

[15]  Nassir Navab,et al.  Concurrent Spatial and Channel Squeeze & Excitation in Fully Convolutional Networks , 2018, MICCAI.

[16]  Seyed-Ahmad Ahmadi,et al.  V-Net: Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation , 2016, 2016 Fourth International Conference on 3D Vision (3DV).

[17]  Xiaogang Wang,et al.  Residual Attention Network for Image Classification , 2017, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[18]  Gang Sun,et al.  Squeeze-and-Excitation Networks , 2017, 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition.

[19]  George Papandreou,et al.  Rethinking Atrous Convolution for Semantic Image Segmentation , 2017, ArXiv.