TY - JOUR
T1 - RCTE
T2 - A reliable and consistent temporal-ensembling framework for semi-supervised segmentation of COVID-19 lesions
AU - Ding, Weiping
AU - Abdel-Basset, Mohamed
AU - Hawash, Hossam
N1 - Funding Information:
This research was supported in part by the National Natural Science Foundation of China under 61300167, 61300169 and 61976120 , supported in part by the Natural Science Foundation of Jiangsu Province under Grant BK20191445, supported in part by the Six Talent Peaks Project of Jiangsu Province under Grant XYDXXJS-048, and sponsored by Qing Lan Project of Jiangsu Province.
Publisher Copyright:
© 2021 Elsevier Inc.
PY - 2021/11
Y1 - 2021/11
N2 - The segmentation of COVID-19 lesions from computed tomography (CT) scans is crucial to develop an efficient automated diagnosis system. Deep learning (DL) has shown success in different segmentation tasks. However, an efficient DL approach requires a large amount of accurately annotated data, which is difficult to aggregate owing to the urgent situation of COVID-19. Inaccurate annotation can easily occur without experts, and segmentation performance is substantially worsened by noisy annotations. Therefore, this study presents a reliable and consistent temporal-ensembling (RCTE) framework for semi-supervised lesion segmentation. A segmentation network is integrated into a teacher-student architecture to segment infection regions from a limited number of annotated CT scans and a large number of unannotated CT scans. The network generates reliable and unreliable targets, and to evenly handle these targets potentially degrades performance. To address this, a reliable teacher-student architecture is introduced, where a reliable teacher network is the exponential moving average (EMA) of a reliable student network that is reliably renovated by restraining the student involvement to EMA when its loss is larger. We also present a noise-aware loss based on improvements to generalized cross-entropy loss to lead the segmentation performance toward noisy annotations. Comprehensive analysis validates the robustness of RCTE over recent cutting-edge semi-supervised segmentation techniques, with a 65.87% Dice score.
AB - The segmentation of COVID-19 lesions from computed tomography (CT) scans is crucial to develop an efficient automated diagnosis system. Deep learning (DL) has shown success in different segmentation tasks. However, an efficient DL approach requires a large amount of accurately annotated data, which is difficult to aggregate owing to the urgent situation of COVID-19. Inaccurate annotation can easily occur without experts, and segmentation performance is substantially worsened by noisy annotations. Therefore, this study presents a reliable and consistent temporal-ensembling (RCTE) framework for semi-supervised lesion segmentation. A segmentation network is integrated into a teacher-student architecture to segment infection regions from a limited number of annotated CT scans and a large number of unannotated CT scans. The network generates reliable and unreliable targets, and to evenly handle these targets potentially degrades performance. To address this, a reliable teacher-student architecture is introduced, where a reliable teacher network is the exponential moving average (EMA) of a reliable student network that is reliably renovated by restraining the student involvement to EMA when its loss is larger. We also present a noise-aware loss based on improvements to generalized cross-entropy loss to lead the segmentation performance toward noisy annotations. Comprehensive analysis validates the robustness of RCTE over recent cutting-edge semi-supervised segmentation techniques, with a 65.87% Dice score.
KW - COVID-19
KW - CT scans
KW - Deep learning
KW - Semi-supervised learning
KW - Temporal-ensembling
UR - http://www.scopus.com/inward/record.url?scp=85111326479&partnerID=8YFLogxK
U2 - 10.1016/j.ins.2021.07.059
DO - 10.1016/j.ins.2021.07.059
M3 - Article
AN - SCOPUS:85111326479
SN - 0020-0255
VL - 578
SP - 559
EP - 573
JO - Information Sciences
JF - Information Sciences
ER -