Supervised Contrastive Learning to Classify Paranasal Anomalies in the Maxillary Sinus

TY - CHAP

T1 - Supervised Contrastive Learning to Classify Paranasal Anomalies in the Maxillary Sinus

AU - Bhattacharya, Debayan

AU - Becker, Benjamin Tobias

AU - Behrendt, Finn

AU - Bengs, Marcel

AU - Beyersdorff, Dirk

AU - Eggert, Dennis

AU - Petersen, Elina

AU - Jansen, Florian

AU - Petersen, Marvin

AU - Cheng, Bastian

AU - Betz, Christian

AU - Schlaefer, Alexander

AU - Hoffmann, Anna Sophie

PY - 2022/9/16

Y1 - 2022/9/16

N2 - Using deep learning techniques, anomalies in the paranasal sinus system can be detected automatically in MRI images and can be further analyzed and classified based on their volume, shape and other parameters like local contrast. However due to limited training data, traditional supervised learning methods often fail to generalize. Existing deep learning methods in paranasal anomaly classification have been used to diagnose at most one anomaly. In our work, we consider three anomalies. Specifically, we employ a 3D CNN to separate maxillary sinus volumes without anomaly from maxillary sinus volumes with anomaly. To learn robust representations from a small labelled dataset, we propose a novel learning paradigm that combines contrastive loss and cross-entropy loss. Particularly, we use a supervised contrastive loss that encourages embeddings of maxillary sinus volumes with and without anomaly to form two distinct clusters while the cross-entropy loss encourages the 3D CNN to maintain its discriminative ability. We report that optimising with both losses is advantageous over optimising with only one loss. We also find that our training strategy leads to label efficiency. With our method, a 3D CNN classifier achieves an AUROC of 0.85 ± 0.03 while a 3D CNN classifier optimised with cross-entropy loss achieves an AUROC of 0.66 ± 0.1.

AB - Using deep learning techniques, anomalies in the paranasal sinus system can be detected automatically in MRI images and can be further analyzed and classified based on their volume, shape and other parameters like local contrast. However due to limited training data, traditional supervised learning methods often fail to generalize. Existing deep learning methods in paranasal anomaly classification have been used to diagnose at most one anomaly. In our work, we consider three anomalies. Specifically, we employ a 3D CNN to separate maxillary sinus volumes without anomaly from maxillary sinus volumes with anomaly. To learn robust representations from a small labelled dataset, we propose a novel learning paradigm that combines contrastive loss and cross-entropy loss. Particularly, we use a supervised contrastive loss that encourages embeddings of maxillary sinus volumes with and without anomaly to form two distinct clusters while the cross-entropy loss encourages the 3D CNN to maintain its discriminative ability. We report that optimising with both losses is advantageous over optimising with only one loss. We also find that our training strategy leads to label efficiency. With our method, a 3D CNN classifier achieves an AUROC of 0.85 ± 0.03 while a 3D CNN classifier optimised with cross-entropy loss achieves an AUROC of 0.66 ± 0.1.

U2 - 10.1007/978-3-031-16437-8_41

DO - 10.1007/978-3-031-16437-8_41

M3 - SCORING: Contribution to collected editions/anthologies

SN - 978-3-031-16436-1

T3 - Lecture Notes in Computer Science

SP - 429

EP - 438

BT - Medical Image Computing and Computer Assisted Intervention – MICCAI 2022

A2 - Wang, Linwei

A2 - Dou, Qi

A2 - Fletcher, P. Thomas

A2 - Speidel, Stefanie

A2 - Li, Shuo

PB - Springer, Cham

CY - Cham

ER -