TY - GEN
T1 - Weakly Supervised Segmentation by a Deep Geodesic Prior
AU - Mortazi, Aliasghar
AU - Khosravan, Naji
AU - Torigian, Drew A.
AU - Kurugol, Sila
AU - Bagci, Ulas
N1 - Funding Information:
A. Mortazi—This work was done partially during internship at Boston Children’s Hospital under the supervision of Dr. Kurugol and was supported partially by Crohns and Colitis Foundation of Americas (CCFA) Career Development Award and AGA-Boston Scientific Technology and Innovation Award.
Publisher Copyright:
© 2019, Springer Nature Switzerland AG.
PY - 2019
Y1 - 2019
N2 - The performance of the state-of-the-art image segmentation methods heavily relies on the high-quality annotations, which are not easily affordable, particularly for medical data. To alleviate this limitation, in this study, we propose a weakly supervised image segmentation method based on a deep geodesic prior. We hypothesize that integration of this prior information can reduce the adverse effects of weak labels in segmentation accuracy. Our proposed algorithm is based on a prior information, extracted from an auto-encoder, trained to map objects’ geodesic maps to their corresponding binary maps. The obtained information is then used as an extra term in the loss function of the segmentor. In order to show efficacy of the proposed strategy, we have experimented segmentation of cardiac substructures with clean and two levels of noisy labels (L1, L2). Our experiments showed that the proposed algorithm boosted the performance of baseline deep learning-based segmentation for both clean and noisy labels by $$4.4\%$$, $$4.6\%$$ (L1), and $$6.3\%$$ (L2) in dice score, respectively. We also showed that the proposed method was more robust in the presence of high-level noise due to the existence of shape priors.
AB - The performance of the state-of-the-art image segmentation methods heavily relies on the high-quality annotations, which are not easily affordable, particularly for medical data. To alleviate this limitation, in this study, we propose a weakly supervised image segmentation method based on a deep geodesic prior. We hypothesize that integration of this prior information can reduce the adverse effects of weak labels in segmentation accuracy. Our proposed algorithm is based on a prior information, extracted from an auto-encoder, trained to map objects’ geodesic maps to their corresponding binary maps. The obtained information is then used as an extra term in the loss function of the segmentor. In order to show efficacy of the proposed strategy, we have experimented segmentation of cardiac substructures with clean and two levels of noisy labels (L1, L2). Our experiments showed that the proposed algorithm boosted the performance of baseline deep learning-based segmentation for both clean and noisy labels by $$4.4\%$$, $$4.6\%$$ (L1), and $$6.3\%$$ (L2) in dice score, respectively. We also showed that the proposed method was more robust in the presence of high-level noise due to the existence of shape priors.
KW - Deep learning
KW - Geodesic prior
KW - Medical image segmentation
KW - Shape prior
KW - Weakly supervised
UR - http://www.scopus.com/inward/record.url?scp=85075671632&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85075671632&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-32692-0_28
DO - 10.1007/978-3-030-32692-0_28
M3 - Conference contribution
AN - SCOPUS:85075671632
SN - 9783030326913
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 238
EP - 246
BT - Machine Learning in Medical Imaging - 10th International Workshop, MLMI 2019, Held in Conjunction with MICCAI 2019, Proceedings
A2 - Suk, Heung-Il
A2 - Liu, Mingxia
A2 - Lian, Chunfeng
A2 - Yan, Pingkun
PB - Springer
T2 - 10th International Workshop on Machine Learning in Medical Imaging, MLMI 2019 held in conjunction with the 22nd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2019
Y2 - 13 October 2019 through 13 October 2019
ER -