대한소아치과학회지 (Journal of the korean academy of Pediatric Dentistry)

  • 제50권3호
  • /
  • Pages.263-276
  • /
  • 2023
  • /
  • 1226-8496(pISSN)
  • /
  • 2288-3819(eISSN)
대한소아치과학회 (Korean Academy of Pediatric Dentistry)
권호 표지
DOI QR코드

DOI QR Code

유치의 치근단 방사선 사진에서 딥 러닝 알고리즘을 이용한 모델의 인접면 우식증 객체 탐지 능력의 평가

Assessment of the Object Detection Ability of Interproximal Caries on Primary Teeth in Periapical Radiographs Using Deep Learning Algorithms

  • 전홍주 (전남대학교 치의학전문대학원 소아치과학교실) ;
  • 김선미 (전남대학교 치의학전문대학원 소아치과학교실) ;
  • 최남기 (전남대학교 치의학전문대학원 소아치과학교실)
  • Hongju Jeon (Department of Pediatric Dentistry, School of Dentistry, Chonnam National University) ;
  • Seonmi Kim (Department of Pediatric Dentistry, School of Dentistry, Chonnam National University) ;
  • Namki Choi (Department of Pediatric Dentistry, School of Dentistry, Chonnam National University)
  • 투고 : 2023.05.11
  • 심사 : 2023.06.25
  • 발행 : 2023.08.31
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

이 연구의 목적은 소아의 치근단 방사선 사진에서 인접면 우식증 객체 탐지 의 객체 탐지를 위해 YOLO (You Only Look Once)를 사용한 모델의 성능을 평가하는 것이다. M6 데이터베이스에서 학습자료군으로 2016개의 치근단 방사선 사진이 선택되었고 이 중 1143개는 한 명의 숙련된 치과의사가 주석 도구를 사용하여 인접면 우식증을 표시하였다. 표시한 주석을 데이터 세트로 변환한 후 단일 합성곱 신경망(CNN) 모델을 기반으로 하는 YOLO를 데이터 세트에 학습시켰다. 187개의 평가자료군에서 객체 탐지 모델 성능 평가를 위해 정확도, 재현율, 특이도, 정밀도, NPV, F1-score, PR 곡선 및 AP를 계산하였다. 결과로 정확도 0.95, 재현율 0.94, 특이도 0.97, 정밀도 0.82, NPV 0.96, F1-score 0.81, AP 0.83으로 인접면 우식증 탐지에 좋은 성능을 보였다. 이 모델은 치과의사에게 치근단 방사선 사진에서 인접면 우식증 병변을 객체 탐지하는 도구로 유용하게 사용될 수 있다.

The purpose of this study was to evaluate the performance of a model using You Only Look Once (YOLO) for object detection of proximal caries in periapical radiographs of children. A total of 2016 periapical radiographs in primary dentition were selected from the M6 database as a learning material group, of which 1143 were labeled as proximal caries by an experienced dentist using an annotation tool. After converting the annotations into a training dataset, YOLO was trained on the dataset using a single convolutional neural network (CNN) model. Accuracy, recall, specificity, precision, negative predictive value (NPV), F1-score, Precision-Recall curve, and AP (area under curve) were calculated for evaluation of the object detection model's performance in the 187 test datasets. The results showed that the CNN-based object detection model performed well in detecting proximal caries, with a diagnostic accuracy of 0.95, a recall of 0.94, a specificity of 0.97, a precision of 0.82, a NPV of 0.96, and an F1-score of 0.81. The AP was 0.83. This model could be a valuable tool for dentists in detecting carious lesions in periapical radiographs.

키워드

참고문헌

  1. Krithiga R, Lakshmi C : A survey: segmentation in dental X-ray images for diagnosis of dental caries. Int J Cont Theory Appl, 9:941-948, 2016.
  2. Selwitz RH, Ismail AI, Pitts NB : Dental caries. Lancet, 369:51-59, 2007. https://doi.org/10.1016/S0140-6736(07)60031-2
  3. Tam LE, McComb D : Diagnosis of occlusal caries: Part II. Recent diagnostic technologies. J Can Dent Assoc, 67:459-463, 2001.
  4. Akkaya N, Kansu O, Kansu H, Cagirankaya LB, Arslan U : Comparing the accuracy of panoramic and intraoral radiography in the diagnosis of proximal caries. Dentomaxillofac Radiol, 35:170-174, 2006. https://doi.org/10.1259/dmfr/26750940
  5. McKnight-Hanes C, Myers DR, Dushku JC, Thompson WO, Durham LC : Radiographic recommendations for the primary dentition: comparison of general dentists and pediatric dentists. Pediatr Dent, 12:212-216, 1990.
  6. da Silva Pierro VS, Barcelos R, de Souza IPR, Raymundo R Jr : Pediatric Bitewing Film Holder: Preschoolers' Acceptance and Radiographs' Diagnostic Quality. Pediatr Dent, 30:342-347, 2008.
  7. Pitts NB : The use of bitewing radiographs in the management of dental caries: scientific and practical considerations. Dentomaxillofac Radio, 25:5-16, 1996. https://doi.org/10.1259/dmfr.25.1.9084279
  8. Gopal KS, Krishnaraj N, Priya M : Faulty radiographs: A retrospective radiographic analysis. Int J Appl Dent Sci, 4:72-76, 2018.
  9. Bini SA : Artificial Intelligence, Machine Learning, Deep Learning, and Cognitive Computing: What Do These Terms Mean and How Will They Impact Health Care? J Arthroplasty, 33:2358-2361, 2018. https://doi.org/10.1016/j.arth.2018.02.067
  10. Lee S, Oh SI, Jo J, Kang S, Shin Y, Park JW : Deep learning for early dental caries detection in bitewing radiographs. Sci Rep, 11:16807, 2021.
  11. Karimi D, Salcudean SE : Reducing the hausdorff distance in medical image segmentation with convolutional neural networks. IEEE Trans Med Imaging, 39:499-513, 2020. https://doi.org/10.1109/TMI.2019.2930068
  12. Redmon J, Divvala S, Girshick R, Farhadi A : You only look once: Unified, real-time object detection. Available from URL: https://www.cv-foundation.org/openaccess/content_cvpr_2016/papers/Redmon_You_Only_Look_CVPR_2016_paper.pdf (Accessed on January 10, 2023).
  13. Yan B, Fan P, Lei X, Liu Z, Yang F : A Real-Time Apple Targets Detection Method for Picking Robot Based on Improved YOLOv5. Remote Sens, 13:1619, 2021.
  14. Girshick R, Donahue J, Darrell T, Malik J : Rich feature hierarchies for accurate object detection and semantic segmentation. Proceedings of the IEEE conference on computer vision and pattern recognition, 580-587, 2014.
  15. Arya C, Tripathi A, Singh P, Diwakar M, Sharma K, Pandey H : Object detection using deep learning: a review. J Phys Conf Series, 1854:012012, 2021.
  16. Kim H, Song JS, Shin TJ, Hyun HK, Kim JW, Jang KT, Kim YJ : Detection of Proximal Caries Lesions with Deep Learning Algorithm. J Korean Acad Pediatr Dent, 49:131-139, 2022. https://doi.org/10.5933/JKAPD.2022.49.2.131
  17. Young DA, Novy BB, Zeller GG, Hale R, Hart TC, Truelove EL; American Dental Association Council on Scientific Affairs : The American Dental Association caries classification system for clinical practice: a report of the American Dental Association Council on Scientific Affairs. J Am Dent Assoc, 146:79-86, 2015. https://doi.org/10.1016/j.adaj.2014.11.018
  18. Tzutalin D : LabelImg. Available from URL: https://github.com/HumanSignal/labelImg (Accessed on October 15, 2022).
  19. Shorten C, Khoshgoftaar TM : A survey on image data augmentation for deep learning. J Big Data, 6:60, 2019.
  20. Rezatofighi H, Tsoi N, Gwak J, Sadeghian A, Reid I, Savarese S : 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, 658-666, 2019.
  21. Ren S, He K, Girshick R, Sun J : Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. IEEE Trans Pattern Anal Mach Intell, 39:1137-1149, 2017. https://doi.org/10.1109/TPAMI.2016.2577031
  22. Park IH, Kim SH : Performance Indicator Survey for Object Detection. Proc. of 20th International Conference on Control, Automation and Systems (ICCAS), 284-288, 2020.
  23. Padilla R, Netto SL, Da Silva EA : A survey on performance metrics for object-detection algorithms. 2020 International Conference on Systems, Signals and Image Processing (IWSSIP). IEEE, 237-242, 2020.
  24. Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, Blondel M, Prettenhofer P, Weiss R, Dubourg V : Scikit-learn: Machine learning in Python. J Mach Learn Res, 12:2825-2830, 2011.
  25. Buitinck L, Louppe G, Blondel M, Pedregosa F, Mueller A, Grisel O, Niculae V, Prettenhofer P, Gramfort A, Grobler J, Layton R, Vanderplas J, Joly A, Holt B, Varoquaux G : API design for machine learning software: experiences from the scikit-learn project. arXiv,1309.0238, 2013.
  26. Krohling LL, de Paula KMP, Behlau M : ROC curve of the Pediatric Voice Related Quality-of-Life Survey (PVRQOL). Codas, 28:311-313, 2016. https://doi.org/10.1590/2317-1782/20162015103
  27. Campbell I : Chi-squared and Fisher-Irwin tests of two-by-two tables with small sample recommendations. Stat Med, 26:3661-3675, 2007. https://doi.org/10.1002/sim.2832
  28. Mo KH, Yoon JH, Kim SG, Lee SH : Detection of proximal caries using laser fluorescence. J Korean Acad Pediatr Dent, 31:323-330, 2004.
  29. Wilson PR, Beynon AD : Mineralization differences between human deciduous and permanent enamel measured by quantitative microradiography. Arch Oral Biol, 34:85-88, 1989. https://doi.org/10.1016/0003-9969(89)90130-1
  30. Seol JH, Oh YH, Lee NY, Lee SH : Detection of early proximal caries with laser fluorescence. J Korean Acad Pediatr Dent, 31:236-246, 2004.
  31. Bayraktar Y, Ayan E : Diagnosis of interproximal caries lesions with deep convolutional neural network in digital bitewing radiographs. Clin Oral Investig, 26:623-632, 2022. https://doi.org/10.1007/s00784-021-04040-1
  32. Srivastava MM, Kumar P, Pradhan L, Varadarajan S : Detection of tooth caries in bitewing radiographs using deep learning. arXiv, 1711.07312, 2017.
  33. Cantu AG, Gehrung S, Krois J, Chaurasia A, Rossi JG, Gaudin R, Elhennawy K, Schwendicke F : Detecting caries lesions of different radiographic extension on bitewings using deep learning. J Dent, 100:103425, 2020.
  34. Lian L, Zhu T, Zhu F, Zhu H : Deep learning for caries detection and classification. Diagnostics, 11:1672, 2021.
  35. Allwright S : What is a good F1 score and how do I interpret it? Available from URL: https://stephenallwright.com/good-f1-score (Accessed on November 8, 2022).
  36. Youssry N, Khattab A : Accurate Real-Time Face Mask Detection Framework Using YOLOv5. 4th IEEE International Conference on Design & Test of Integrated Micro & Nano-Systems, DTS2022, 2022.