R2U3D: Recurrent Residual 3D U-Net for Lung Segmentation
Document Type
Article
Publication Date
6-16-2021
Publication Source
IEEE Access
Abstract
3D Lung segmentation is essential since it processes the volumetric information of the lungs, removes the unnecessary areas of the scan, and segments the actual area of the lungs in a 3D volume. Recently, the deep learning model, such as U-Net outperforms other network architectures for biomedical image segmentation. In this paper, we propose a novel model, namely, Recurrent Residual 3D U-Net (R 2 U3D), for the 3D lung segmentation task. In particular, the proposed model integrates 3D convolution into the Recurrent Residual Neural Network based on U-Net. It helps learn spatial dependencies in 3D and increases the propagation of 3D volumetric information. The proposed R 2 U3D network is trained on the publicly available dataset LUNA16 and it achieves state-of-the-art performance on both LUNA16 (testing set) and VESSEL12 dataset. In addition, we show that training the R 2 U3D model with a smaller number of CT scans, i.e., 100 scans, without applying data augmentation achieves an outstanding result in terms of Soft Dice Similarity Coefficient (Soft-DSC) of 0.9920.
Inclusive pages
88835 - 88843
ISBN/ISSN
2169-3536
Publisher
IEEE
Volume
9
Keywords
Lung, Feature extraction, Three-dimensional displays, Image segmentation, Deep learning, Computed tomography, Biomedical imaging, 3D Lung segmentation, R²U3D, semantic segmentation, deep CNN, biomedical image analysis
eCommons Citation
Kadia, Dhaval D.; Alom, Md Zahangir; Burada, Ranga; Nguyen, Tam V.; and Asari, Vijayan K., "R2U3D: Recurrent Residual 3D U-Net for Lung Segmentation" (2021). Computer Science Faculty Publications. 232.
https://ecommons.udayton.edu/cps_fac_pub/232
COinS
