3D-2D distance maps conversion enhances classification of craniosynostosis

• Verfasst von: Schaufelberger, Matthias [VerfasserIn]
• Verfasst von: Kaiser, Christian [VerfasserIn]
• Verfasst von: Kühle, Reinald [VerfasserIn]
• Verfasst von: Wachter, Andreas [VerfasserIn]
• Verfasst von: Bouffleur, Frederic [VerfasserIn]
• Verfasst von: Hagen, Niclas [VerfasserIn]
• Verfasst von: Ringwald, Friedemann [VerfasserIn]
• Verfasst von: Eisenmann, Urs [VerfasserIn]
• Verfasst von: Hoffmann, Jürgen [VerfasserIn]
• Verfasst von: Engel, Michael [VerfasserIn]
• Verfasst von: Freudlsperger, Christian [VerfasserIn]
• Verfasst von: Nahm, Werner [VerfasserIn]
• Titel: 3D-2D distance maps conversion enhances classification of craniosynostosis
• Verf.angabe: Matthias Schaufelberger, Christian Kaiser, Reinald Kühle, Andreas Wachter, Frederic Weichel, Niclas Hagen, Friedemann Ringwald, Urs Eisenmann, Jürgen Hoffmann, Michael Engel, Christian Freudlsperger, and Werner Nahm
• E-Jahr: 2023
• Jahr: November 2023
• Umfang: 10 S.
• Illustrationen: Illustrationen
• Fussnoten: Gesehen am 06.02.2024
• Titel Quelle: Enthalten in: Institute of Electrical and Electronics EngineersIEEE transactions on biomedical engineering
• Ort Quelle: New York, NY : IEEE, 1964
• Jahr Quelle: 2023
• Band/Heft Quelle: 70(2023), 11 vom: Nov., Seite 3156-3165
• ISSN Quelle: 1558-2531
• DOI: doi:10.1109/TBME.2023.3278030
• Datenträger: Online-Ressource
• Sprache: eng
• Sach-SW: 2D conversion
• Sach-SW: classification
• Sach-SW: CNN
• Sach-SW: convolutional neural network
• Sach-SW: craniosynostosis
• Sach-SW: data augmentation
• Sach-SW: distance map
• Sach-SW: Head
• Sach-SW: Magnetic heads
• Sach-SW: Pediatrics
• Sach-SW: photogrammetric surface scans
• Sach-SW: resolution
• Sach-SW: Shape
• Sach-SW: Surgery
• Sach-SW: Three-dimensional displays
• Sach-SW: Training
• K10plus-PPN: 1880051370

Abstract

Objective: Diagnosis of craniosynostosis using photogrammetric 3D surface scans is a promising radiation-free alternative to traditional computed tomography. We propose a 3D surface scan to 2D distance map conversion enabling the usage of the first convolutional neural networks (CNNs)-based classification of craniosynostosis. Benefits of using 2D images include preserving patient anonymity, enabling data augmentation during training, and a strong under-sampling of the 3D surface with good classification performance. Methods: The proposed distance maps sample 2D images from 3D surface scans using a coordinate transformation, ray casting, and distance extraction. We introduce a CNN-based classification pipeline and compare our classifier to alternative approaches on a dataset of 496 patients. We investigate into low-resolution sampling, data augmentation, and attribution mapping. Results: Resnet18 outperformed alternative classifiers on our dataset with an F1-score of 0.964 and an accuracy of 98.4%. Data augmentation on 2D distance maps increased performance for all classifiers. Under-sampling allowed 256-fold computation reduction during ray casting while retaining an F1-score of 0.92. Attribution maps showed high amplitudes on the frontal head. Conclusion: We demonstrated a versatile mapping approach to extract a 2D distance map from the 3D head geometry increasing classification performance, enabling data augmentation during training on 2D distance maps, and the usage of CNNs. We found that low-resolution images were sufficient for a good classification performance. Significance: Photogrammetric surface scans are a suitable craniosynostosis diagnosis tool for clinical practice. Domain transfer to computed tomography seems likely and can further contribute to reducing ionizing radiation exposure for infants.