FU-Net: Multi-class Image Segmentation Using Feedback Weighted U-Net

Jafari, Mina; Li, Ruizhe; Xing, Yue; Auer, Dorothee; Francis, Susan; Garibaldi, Jonathan; Chen, Xin

Abstract

© 2019, Springer Nature Switzerland AG. In this paper, we present a generic deep convolutional neural network (DCNN) for multi-class image segmentation. It is based on a well-established supervised end-to-end DCNN model, known as U-net. U-net is firstly modified by adding widely used batch normalization and residual block (named as BRU-net) to improve the efficiency of model training. Based on BRU-net, we further introduce a dynamically weighted cross-entropy loss function. The weighting scheme is calculated based on the pixel-wise prediction accuracy during the training process. Assigning higher weights to pixels with lower segmentation accuracies enables the network to learn more from poorly predicted image regions. Our method is named as feedback weighted U-net (FU-net). We have evaluated our method based on T1-weighted brain MRI for the segmentation of midbrain and substantia nigra, where the number of pixels in each class is extremely unbalanced to each other. Based on the dice coefficient measurement, our proposed FU-net has outperformed BRU-net and U-net with statistical significance, especially when only a small number of training examples are available. The code is publicly available in GitHub (GitHub link: https://github.com/MinaJf/FU-net).

Citation

Jafari, M., Li, R., Xing, Y., Auer, D., Francis, S., Garibaldi, J., & Chen, X. (2019). FU-Net: Multi-class Image Segmentation Using Feedback Weighted U-Net. In Image and Graphics: 10th International Conference, ICIG 2019, Beijing, China, August 23–25, 2019, Proceedings, Part II (529-537). Springer Verlag. https://doi.org/10.1007/978-3-030-34110-7_44

• Online Publication Date: Nov 28, 2019
• Publication Date: 2019
• Deposit Date: Jan 10, 2020
• Publicly Available Date: Nov 29, 2020
• Publisher: Springer Verlag
• Pages: 529-537
• Series Title: Lecture Notes in Computer Science
• Series Number: 11902
• Book Title: Image and Graphics: 10th International Conference, ICIG 2019, Beijing, China, August 23–25, 2019, Proceedings, Part II
• ISBN: 9783030341091; 9783030341107
• DOI: https://doi.org/10.1007/978-3-030-34110-7_44
• Public URL: https://nottingham-repository.worktribe.com/output/3518051
• Additional Information: First Online: 28 November 2019; Conference Acronym: ICIG; Conference Name: International Conference on Image and Graphics; Conference City: Beijing; Conference Country: China; Conference Year: 2019; Conference Start Date: 23 August 2019; Conference End Date: 25 August 2019; Conference Number: 10; Conference ID: icig2019; Conference URL: http://www.csig.org.cn/detail/2669