실무 기반 3D 치아 스캔 데이터 세분화를 위한 딥러닝 모델 성능 평가: MeshSegNet과 DilatedToothSegNet 모델 활용
(Performance Evaluation of Deep Learning Models for Practical 3D Tooth Scan Data Segmentation: Application of MeshSegNet and DilatedToothSegNet Models)
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
본 연구는 치과 진료 현장에서 빈번히 발생하는 작업치(크라운, 임플란트, 인레이)와 국소 스캔 데이터를 포함한 910건의 3D 치아 스캔 데이터를 수집, 15개의 클래스(치은 및 치아 번호별)로 라벨링하여 임상 적용 가능성을 높이는 것을 목표로 한다. 기존 연구는 전체 구강 구조 데이터를 활용해 실제 임상 데이터의 불규칙성을 충분히 반영하지 못했으나, 본 연구는 이를 개선하고자 딥러닝 모델인 MeshSegNet과 DilatedToothSegNet을 활용해 치아 번호별 3D 세그멘테이션 성능을 비교·평가하였다. 해상도 별 비교를 위해 메쉬 셀을 10,000개와 16,000개로 다운샘플링하여 학습시켰으며, DSC와 mIoU를 사용해 성능을 측정했다. DilatedToothSegNet(16,000 Cells)은 상악과 하악 데이터에서 MeshSegNet 대비 DSC 및 mIoU 지표 모두 약 2.2~3.5% 우수한 성능을 보였으며, 작업치를 포함한 국소 스캔 데이터에서도 안정적이고 일관된 세분화 성능을 나타냈다.
Abstract
This study aims to improve the clinical applicability of 3D segmentation models for tooth numbering by collecting and labeling 910 3D tooth scans, including prosthetic (crown, implant, inlay) and localized scan data, into 15 classes (gingiva and individual tooth numbers). Previous studies often relied on full oral structure data, failing to capture the irregularities of real clinical scenarios. To address this limitation, this research compares the performance of two deep learning models, MeshSegNet and DilatedToothSegNet, for tooth-numbered 3D segmentation. The models were trained with mesh cell resolutions of 10,000 and 16,000 to assess accuracy at different levels. Evaluation metrics, including Dice Similarity Coefficient (DSC) and mean Intersection over Union (mIoU), were used to measure performance. Results showed that the DilatedToothSegNet model with 16,000 cells outperformed MeshSegNet, achieving a 2.2–3.5% improvement in both DSC and mIoU metrics for maxillary and mandibular data. Additionally, it demonstrated consistent and reliable performance on localized scans, including prosthetic data.
참고문헌 (22)
- Z. S. Haidar, "Digital Dentistry: Past, Present, and Future", Digital Medicine and Healthcare Technology, 2 (2023)
- 이가영, "우리나라 국민의 외래 치과의료비 지출 경향 : 2011-2017 한국의료패널 분석", 치과의료정책연구원, (22) (2020)
- 건강보험심사평가원, "심사평가원, ‘치과 외래 진료현황 분석’ 발표", ~ (2023)
- 김려운, "치과 디지털 인상의 이해와 적용", 대한치과재료학회지, 41(4), pp.253~261 (2014)
- C. Lian, "Deep Multi-Scale Mesh Feature Learning for Automated Labeling of Raw Dental Surfaces from 3D Intraoral Scanners", IEEE Transactions on Medical Imaging, 39(7), pp.2440~2450 (2020)
- J. -S. An, "Performance Comparison of 3D Segmentation Models Based on Point Cloud and Mesh for Dental Scan Data", KIISE Transactions on Computing Practices, 29(4), pp.158~165 (2023)
- L. Krenmayr, "DilatedToothSegNet : Tooth Segmentation Network on 3D Dental Meshes Through Increasing Receptive Vision", Journal of Imaging and Information Medicine, 37(4), pp.1846~1862 (2024)
- 장윤정, "Digital intraoral impression for immediate provisional restoration of maxillary single implant : A case report", Journal of Korean Dental Prosthodontics, 53(3), pp.234~243 (2015)
- Hong Tai Tran, "Liver Segmentation and 3D Modeling from Abdominal CT Images", Smart Media Journal, 5(1), pp.1~6 (2016)
- M. A. Balafar, "Review of Brain MRI Image Segmentation Methods", Artificial Intelligence Review, 33(3), pp.261~274 (2010)
- Ö. Çiçek, "3D U-Net : Learning Dense Volumetric Segmentation from Sparse Annotation", Proceedings of the 19th International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI 2016) (2016)
- F. Milletari, "V-Net : Fully Convolutional Neural Networks for Volumetric Medical Image Segmentation", Proceedings of the 2016 Fourth International Conference on 3D Vision (3DV) (2016)
- Vo Thi Tuong Vi, "Automatic Extraction of Liver Region from Medical Images by Using an MFUnet", Smart Media Journal, 9(3), pp.59~70 (2020)
- Afnan Ghafoor, "Attention-based deep learning framework for skin lesion segmentation", Smart Media Journal, 13(3), pp.53~61 (2024)
- C. R. Qi, "PointNet : Deep Learning on Point Sets for 3D Classification and Segmentation", Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition(CVPR) (2017)
- T. -H. Wu, "Machine Learning in Dentistry" (2021), Springer
- "Mesh_Labeler Github" (2021)[URL]
- "MeshSegNet Github" (2020)[URL]
- "DilatedToothSegNet Github" (2024)[URL]
- L. R. Dice, "Measures of the Amount of Ecologic Association Between Species", Ecology, 26(3), pp.297~302 (1945)
- M. Everingham, "The Pascal Visual Object Classes(VOC)Challenge", International Journal of Computer Vision, 88(2), pp.303~338 (2010)
- F. Yi, "Image Segmentation: A Survey of Graph-Cut Methods", Proceedings of the 2012 International Conference on Systems and Informatics (ICSAI2012) (2012)