[논문]자동 치아 분할용 종단 간 시스템 개발을 위한 선결 연구: 딥러닝 기반 기준점 설정 알고리즘

서경덕; 이세나; 진용규; 양세정

doi:10.9718/jber.2023.44.5.346

자동 치아 분할용 종단 간 시스템 개발을 위한 선결 연구: 딥러닝 기반 기준점 설정 알고리즘
Prerequisite Research for the Development of an End-to-End System for Automatic Tooth Segmentation: A Deep Learning-Based Reference Point Setting Algorithm 원문보기

Journal of biomedical engineering research : the official journal of the Korean Society of Medical & Biological Engineering, v.44 no.5, 2023년, pp.346 - 353

서경덕 (연세대학교 의공학과) , 이세나 (연세대학교 원주의과대학 정밀의학과) , 진용규 (주식회사 디오코) , 양세정 (연세대학교 원주의과대학 정밀의학과)

Abstract ▼ AI-Helper

In this paper, we propose an innovative approach that leverages deep learning to find optimal reference points for achieving precise tooth segmentation in three-dimensional tooth point cloud data. A dataset consisting of 350 aligned maxillary and mandibular cloud data was used as input, and both end coordinates of individual teeth were used as correct answers. A two-dimensional image was created by projecting the rendered point cloud data along the Z-axis, where an image of individual teeth was created using an object detection algorithm. The proposed algorithm is designed by adding various modules to the Unet model that allow effective learning of a narrow range, and detects both end points of the tooth using the generated tooth image. In the evaluation using DSC, Euclid distance, and MAE as indicators, we achieved superior performance compared to other Unet-based models. In future research, we will develop an algorithm to find the reference point of the point cloud by back-projecting the reference point detected in the image in three dimensions, and based on this, we will develop an algorithm to divide the teeth individually in the point cloud through image processing techniques.

주제어

표/그림 (10)

표 표 1. 치아 구성에 따른 유형별 데이터 정보 Table 1. Data information by type according to tooth composition
그림 그림 1. 데이터 가공 및 생성 시스템 모식도 Fig. 1. Overview of the Data Generation Systems
그림 그림 2. 기준점 설정 알고리즘의 개요 Fig. 2. Overview of the Reference Point Setting Algorithm
그림 그림 3. 좌표정보층과 특징맵의 결합 Fig. 3. Combination of Coordinate Layers and Feature map
그림 그림 4. 제안하는 모델의 구조도 Fig. 4. Structure of the Proposed Network
그림 그림 5. Residual Shortcut의 모식도 Fig. 5. Residual Shortcut
표 표 2. 모듈 유무에 따른 결과 Table 2. Evaluation results based on the presence of modules
표 표 3. 모델 비교 결과 Table 3. Comparison of model results
표 그림 6. 모듈 유무에 따른 결과 시각화, (a) Baseline(Unet)의 결과, (b) (a) + 좌표정보층의 결과, (c) (b) + Attention 모듈의 결과, (d) 제안하는 모델의 결과(화살표: 차이가 특히 큰 부분) Fig. 6. Visualization of results with or without modules, (a) Results of Baseline (Unet), (b) (a) + Results of Coordinate Layers, (c) (b) + Results of Attention Module, (d) Results of the proposed model(Red arrows: where the difference is particularly large)
그림 그림 7. 히트맵 활성 범위에 대한 Jensen-Shannon Divergence 규제의 영향, a) 유클리드 손실함수, b) 제안하는 손실함수 Fig. 7. The impact of Jensen-Shannon Divergence Regularization on the heatmap activation range: a) Euclidean loss, b) Proposed loss

참고문헌 (24)

Jang KA, Heo SE, Kang HK, Lee SJ. A Convergence Study on the Changes of Awareness and Preference according to the Clinical Application Experience of Digital Intraoral Scanners in Dental Hygienists. Journal of the Korea Convergence Society. 2018;9(7):135-140.？
Yoon-Ji Kim, Hye-Jung Ha, Sung-Jong Lee, Eon-Hwa Lee, and Jae-Jun Ryu. Digital Orthodontics using Customized Appliance System. The Journal Of The Korean Dental Association. 2016;54(2):134-141.？
Yoon-Ji Kim, Myung-Su Kim, and Dong-Yul Lee. Clinical Effectiveness and Limitations of Digital Orthodontic Appliances. Clinical Journal of Korean Association of Orthodontists. 2019;9(3):202-216.？

상세보기
Kweon-soo Seo, Sunjai Kim, Joo Hyun Kwon, and chang jae seung. Implant Digital Impression with Intraoral Scanners: A Literature Review. The Journal of Korean Academy of Oral & Maxillofacial Implantology. 2017;21(1):2-13.？
Jiman Park, Koak, Jai Young, Kwon, Hobeom, Kim, Myung-Joo, Seong Kyun Kim, Sung-Hun Kim, Yeo, In-Sung, Lim, Youngjun, Han, Jung-Suk and Heo, SeongJoo. Implant-supported Restoration Cases Fabricated from Digital Impression Data with the Help of Intraoral Scanner and Virtual Articulator. The Journal of Korean Academy of Oral & Maxillofacial Implantology. 2017;21(1):14-23.？
Andrews LF. The six keys to normal occlusion. Am J Orthod. 1972;62(3):296-309.？

상세보기
Miethke, Reggie, Lindenau, S, Dietrich, K. The effect of Frankel's function regulator type III on the apical base. European journal of orthodontics. 2003;25:311-318.？

상세보기
GUO, Yulan, et al. Deep learning for 3d point clouds: A survey. IEEE transactions on pattern analysis and machine intelligence. 2020;43(12):4338-4364.？
Charles RQ, Su H, Kaichun M, Guibas LJ. PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2017;77-85.？
Charles R Qi, Li Yi, Hao Su, and Leonidas J. Guibas. PointNet++: deep hierarchical feature learning on point sets in a metric space. In Proceedings of the 31st International Conference on Neural Information Processing Systems (NIPS'17). 2017;5105-5114.？
Wang Yue, et al. Dynamic graph cnn for learning on point clouds. ACM Transactions on Graphics (tog), 2019;38(5):1-12.？
Jana, Ananya & Maiti, Aniruddha & Metaxas, Dimitris. A Critical Analysis of the Limitation of Deep Learning based 3D Dental Mesh Segmentation Methods in Segmenting Partial Scans. 2023.？
Gomez de Mariscal, Estibaliz & Maska, Martin & Kotrbova, Anna & Pospichalova, Vendula & Matula, Pavel & Munoz-Barrutia, Arrate. Deep-Learning-Based Segmentation of Small Extracellular Vesicles in Transmission Electron Microscopy Images. Scientific Reports. 2019;9:1-10.？
Falk T, et al. U-Net: Deep learning for cell counting detection and morphometry. Nature Methods. 2018;16:67-70.？
Zhou Zongwei, et al. Unet++: Redesigning skip connections to exploit multiscale features in image segmentation. IEEE transactions on medical imaging. 2019;39(6):1856-1867.？
Huang Huimin, et al. Unet 3+: A full-scale connected unet for medical image segmentation. In: ICASSP 2020-2020 IEEE international conference on acoustics, speech and signal processing (ICASSP). 2020; 1055-1059.？
Rosanne Liu, Joel Lehman, Piero Molino, Felipe Petroski Such, Eric Frank, Alex Sergeev, and Jason Yosinski. An intriguing failing of convolutional neural networks and the CoordConv solution. In Proceedings of the 32nd International Conference on Neural Information Processing Systems (NIPS'18). 2018;9628-9639.？
Terven JR, Cordova-Esparza D. A comprehensive review of YOLO: From YOLOv1 to YOLOv8 and beyond. arXiv. 2023;arXiv:2304.00501.？
Botchkarev, Alexei. Performance Metrics (Error Measures) in Machine Learning Regression, Forecasting and Prognostics: Properties and Typology. Interdisciplinary Journal of Information, Knowledge, and Management. 2019;14:45-79？
Dokmanic Ivan, et al. Euclidean distance matrices: essential theory, algorithms, and applications. IEEE Signal Processing Magazine, 2015;32(6):12-30.？

상세보기
Menendez ML, et al. The jensen-shannon divergence. Journal of the Franklin Institute. 1997;334(2):307-318.？

상세보기
Englesson Erik, Azizpour Hossein. Generalized jensen-shannon divergence loss for learning with noisy labels. Advances in Neural Information Processing Systems. 2021;34:30284-30297.？
Zou KH, Warfield SK, Bharatha A, Tempany CM, Kaus MR, Haker SJ, Wells WM, 3rd, Jolesz FA, Kikinis R. Statistical validation of image segmentation quality based on a spatial overlap index. Academic radiology. 2004;11(2):178-189.？
Rezatofighi Hamid, et al. Generalized intersection over union: A metric and a loss for bounding box regression. Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2019;658-666.

내보내기 구분	파일저장 인쇄 메일전송
구성항목	기본정보 상세정보 관리번호, 논문명, 저널/프로시딩명, 저자 , 발행년, 권, 호, 시작페이지, 끝페이지, 발행기관 관리번호, 논문명, 대등논문명, 저자 , 저널/프로시딩명, 발행기관, 발행년, 발행언어, 권, 호, 시작페이지, 끝페이지, ISBN, ISSN, 주제분야, 키워드, 초록(한글), 초록(영문), 저자(소속기관)
저장형식	Text(ASCII format) Excel format RefWorks Direct Export RIS format (for Reference Manager, ProCite, EndNote), Scholar's Aids, Mendeley
메일정보	받는사람 (필수) @ 보내는사람 (선택) @ 제목 내용 KISTI 검색결과 이메일 서비스
안내	총 건의 자료가 검색되었습니다. 다운받으실 자료의 인덱스를 입력하세요. (1-10,000) 검색결과의 순서대로 최대 10,000건 까지 다운로드가 가능합니다. 데이타가 많을 경우 속도가 느려질 수 있습니다.(최대 2~3분 소요) 다운로드 파일은 UTF-8 형태로 저장됩니다. 파일의 내용이 제대로 보이지 않을실 때는 웹브라우저 상단의 보기 -> 인코딩 -> 자동선택 여부를 확인하십시오. ~ Text(ASCII format) Excel format

연합인증