TY - GEN
T1 - Exigent Examiner and Mean Teacher: An Advanced 3D CNN-Based Semi-Supervised Brain Tumor Segmentation Framework
AU - Wang, Ziyang
AU - Voiculescu, Irina
PY - 2023/10/8
Y1 - 2023/10/8
N2 - With the rise of deep learning applications to medical imaging, there has been a growing appetite for large and well-annotated datasets, yet annotation is time-consuming and hard to come by. In this work, we train a 3D semantic segmentation model in an advanced semi-supervised learning fashion. The proposed SSL framework consists of three models: a Student model that learns from annotated data and a large amount of raw data, a Teacher model with the same architecture as the student, updated by self-ensembling and which supervises the student through pseudo-labels, and an Examiner model that assesses the quality of the student’s inferences. All three models are built with 3D convolutional operations. The overall framework mimics a collaboration between a consistency training Student ↔ Teacher module and an adversarial training Examiner ↔ Student module. The proposed method is validated with various evaluation metrics on a public benchmarking 3D MRI brain tumor segmentation dataset. The experimental results of the proposed method outperform pre-existing semi-supervised methods. The source code, baseline methods, and dataset are available at https://github.com/ziyangwang007/CV-SSL-MIS.
AB - With the rise of deep learning applications to medical imaging, there has been a growing appetite for large and well-annotated datasets, yet annotation is time-consuming and hard to come by. In this work, we train a 3D semantic segmentation model in an advanced semi-supervised learning fashion. The proposed SSL framework consists of three models: a Student model that learns from annotated data and a large amount of raw data, a Teacher model with the same architecture as the student, updated by self-ensembling and which supervises the student through pseudo-labels, and an Examiner model that assesses the quality of the student’s inferences. All three models are built with 3D convolutional operations. The overall framework mimics a collaboration between a consistency training Student ↔ Teacher module and an adversarial training Examiner ↔ Student module. The proposed method is validated with various evaluation metrics on a public benchmarking 3D MRI brain tumor segmentation dataset. The experimental results of the proposed method outperform pre-existing semi-supervised methods. The source code, baseline methods, and dataset are available at https://github.com/ziyangwang007/CV-SSL-MIS.
KW - Adversarial Training
KW - Brain Tumour Segmentation
KW - Image Semantic Segmentation
KW - Mean Teacher
KW - Semi-Supervised Learning
UR - http://www.scopus.com/inward/record.url?scp=85174703607&partnerID=8YFLogxK
UR - https://link.springer.com/chapter/10.1007/978-3-031-44917-8_17
U2 - 10.1007/978-3-031-44917-8_17
DO - 10.1007/978-3-031-44917-8_17
M3 - Conference publication
AN - SCOPUS:85174703607
SN - 9783031471964
VL - 14307
T3 - Lecture Notes in Computer Science (LNCS)
SP - 181
EP - 190
BT - Medical Image Learning with Limited and Noisy Data.
A2 - Xue, Zhiyun
A2 - Antani, Sameer
A2 - Zamzmi, Ghada
A2 - Yang, Feng
A2 - Rajaraman, Sivaramakrishnan
A2 - Huang, Sharon Xiaolei
A2 - Linguraru, Marius George
A2 - Liang, Zhaohui
PB - Springer
T2 - 2nd Workshop on Medical Image Learning with Noisy and Limited Data, MILLanD 2023
Y2 - 8 October 2023 through 8 October 2023
ER -