Acta Photonica Sinica, Volume. 50, Issue 2, 65(2021)
Segmentation of Lung Nodules in CT Images Using Improved U-Net++
Figures & Tables(13)
Fig. 1. Comparison of multi-level feature encoding-decoding structure
Fig. 6. Comparison of segmentation results of pulmonary nodules in LIDC dataset
Fig. 7. Loss-DICE curves of
Fig. 8. Comparison of DICE change curves during network training in LIDC dataset
Fig. 9. Comparison of segmentation results of pulmonary nodules in CQUCH-LC dataset
Fig. 10. Loss-DICE curves of
Fig. 11. Comparison of DICE change curves during network training in CQUCH-LC dataset
Table 1. Segmentation results of multiple metrics on the testing set of LIDC dataset by different algorithms
View tableView in ArticleTable 1. Segmentation results of multiple metrics on the testing set of LIDC dataset by different algorithms
Algorithm IoU DICE Sensitivity Precision U-Net 82.03 87.64 89.76 89.91 U-Net++ w/ DS 84.24 88.96 90.46 91.10 U-Net++ w/o DS 83.17 88.50 89.50 90.90 iU-Net++ 87.40 90.83 91.94 92.17
Table 2. Segmentation results of multiple metrics on the testing set of CQUCH-LC dataset by different algorithms
View tableView in ArticleTable 2. Segmentation results of multiple metrics on the testing set of CQUCH-LC dataset by different algorithms
Algorithm IoU DICE Sensitivity Precision U-Net 75.00 83.86 87.37 84.21 U-Net++ w/ DS 78.31 86.09 87.37 88.31 U-Net++ w/o DS 77.97 86.22 87.70 87.19 iU-Net++ 80.59 88.23 89.15 89.11
Tools
Get Citation
Copy Citation Text
Hong HUANG, Rongfei LÜ, Junli TAO, Yuan LI, Jiuquan ZHANG. Segmentation of Lung Nodules in CT Images Using Improved U-Net++[J]. Acta Photonica Sinica, 2021, 50(2): 65
Paper Information
Category: Image Processing
Received: --
Accepted: --
Published Online: Aug. 26, 2021
The Author Email:
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20