TY - GEN
T1 - An Ensemble of Deep Neural Networks for Segmentation of Lung and Clavicle on Chest Radiographs
AU - Zhang, Jingyi
AU - Xia, Yong
AU - Zhang, Yanning
N1 - Publisher Copyright:
© 2020, Springer Nature Switzerland AG.
PY - 2020
Y1 - 2020
N2 - Accurate segmentation of lungs and clavicles on chest radiographs plays a pivotal role in screening, diagnosis, treatment planning, and prognosis of many chest diseases. Although a number of solutions have been proposed, both segmentation tasks remain challenging. In this paper, we propose an ensemble of deep segmentation models (enDeepSeg) that combines the U-Net and DeepLabv3+ to address this challenge. We first extract image patches to train the U-Net and DeepLabv3+ model, respectively, and then use the weighted sum of the segmentation probability maps produced by both models to determine the label of each pixel. The weight of each model is adaptively estimated according to its error rate on the validation set. We evaluated the proposed enDeepSeg model on the Japanese Society of Radiological Technology (JSRT) database and achieved an average Jaccard similarity coefficient (JSC) of 0.961 and 0.883 in the segmentation of lungs and clavicles, respectively, which are higher than those obtained by ten lung segmentation and six clavicle segmentation algorithms. Our results suggest that the enDeepSeg model is able to segment lungs and clavicles on chest radiographs with the state-of-the-art accuracy.
AB - Accurate segmentation of lungs and clavicles on chest radiographs plays a pivotal role in screening, diagnosis, treatment planning, and prognosis of many chest diseases. Although a number of solutions have been proposed, both segmentation tasks remain challenging. In this paper, we propose an ensemble of deep segmentation models (enDeepSeg) that combines the U-Net and DeepLabv3+ to address this challenge. We first extract image patches to train the U-Net and DeepLabv3+ model, respectively, and then use the weighted sum of the segmentation probability maps produced by both models to determine the label of each pixel. The weight of each model is adaptively estimated according to its error rate on the validation set. We evaluated the proposed enDeepSeg model on the Japanese Society of Radiological Technology (JSRT) database and achieved an average Jaccard similarity coefficient (JSC) of 0.961 and 0.883 in the segmentation of lungs and clavicles, respectively, which are higher than those obtained by ten lung segmentation and six clavicle segmentation algorithms. Our results suggest that the enDeepSeg model is able to segment lungs and clavicles on chest radiographs with the state-of-the-art accuracy.
KW - Chest radiograph
KW - Clavicle segmentation
KW - Deep learning
KW - Lung segmentation
UR - http://www.scopus.com/inward/record.url?scp=85079095422&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-39343-4_38
DO - 10.1007/978-3-030-39343-4_38
M3 - 会议稿件
AN - SCOPUS:85079095422
SN - 9783030393427
T3 - Communications in Computer and Information Science
SP - 450
EP - 458
BT - Medical Image Understanding and Analysis - 23rd Conference, MIUA 2019, Proceedings
A2 - Zheng, Yalin
A2 - Williams, Bryan M.
A2 - Chen, Ke
PB - Springer
T2 - 23rd Conference on Medical Image Understanding and Analysis, MIUA 2019
Y2 - 24 July 2019 through 26 July 2019
ER -