Yongjia Huang, Zaifeng Shi, Zhongqi Wang, Zhe Wang. Improved U-Net Based on Mixed Loss Function for Liver Medical Image Segmentation[J]. Laser & Optoelectronics Progress, 2020, 57(22): 221003
- Laser & Optoelectronics Progress
- Vol. 57, Issue 22, 221003 (2020)
Fig. 1. Structure of proposed network
Fig. 2. Improved U-Net structure
Fig. 3. Residual refine module of network structure. (a) General residual refine module; (b) improved residual refine module
Fig. 4. Training error and test accuracy of improved U-Net. (a) Liver tumor segmentation; (b) liver segmentation
Fig. 5. Segmentation results of liver images obtained by different networks
Fig. 6. Box plot of Dice coefficient of liver segmentation
Fig. 7. Segmentation results of liver tumor images obtained by different networks
Fig. 8. Box plot of Dice coefficient of liver tumor segmentation
Fig. 9. Segmentation results of big nodules
Fig. 10. Segmentation results of small nodules
|
Table 1. Performance comparison of different networks for liver image segmentation
|
Table 2. Performance comparison of different networks for liver tumor image segmentation
|
Table 3. Performance comparison of different networks
Set citation alerts for the article
Please enter your email address
© Copyright 2018-2021 | Chinese Laser Press. All Rights Reserved 沪ICP备15018463号-20