IADR Abstract Archives
Digital aid in Diagnosis of periodontitis by deep learning alogrithim
Objectives: - The proposed AI system leverages deep learning algorithms to automate the analysis of periapical radiographs.
- detection and measurement of bone loss in a peripapical radiograph
- detection of furcation involvement
- ultimately providing the ai model with a full mouth series of radiographs and achieving diagnosis for the periodontal status of the patient
Methods: Periapical radiographs were collected from patients attending university of Jordan specialist’s and student’s dental clinics, 5600 periapical radiographs were acquired, and further analysis was done to exclude radiographs of bad quality resulting in 4800 radiographs. These radiographs were further analyzed and uploaded on a platform for annotation; bone, crown and root in each radiograph were labeled separately. 1300 radiograph were used for furcation involvement detection, furcations on each radiograph were labeled. Finally, these radiographs were further processed using DensNet and ResNet for bone loss identification and furcation identification, and after identification of bone loss, image analysis was done to predict the amount of bone loss in millimeters.
Results: Accuracy for peri apical radiograph segmentation and bone loss measurement is 92%
accuracy for furcation involvement detection in upper molars is 50%
accuracy for furcation involvement detection in lower molars is 90%
Conclusions: the application of artificial intelligence in detecting and measuring bone loss in periapical radiographs, along with furcation involvement identification, represents a significant advancement in dental radiology. This automated approach provides precise and reliable assessments, assisting clinicians in diagnosing and managing periodontal disease more effectively. The integration of AI technology in dental practices holds great potential for improving patient care and outcomes in the field of periodontology.
- detection and measurement of bone loss in a peripapical radiograph
- detection of furcation involvement
- ultimately providing the ai model with a full mouth series of radiographs and achieving diagnosis for the periodontal status of the patient
Methods: Periapical radiographs were collected from patients attending university of Jordan specialist’s and student’s dental clinics, 5600 periapical radiographs were acquired, and further analysis was done to exclude radiographs of bad quality resulting in 4800 radiographs. These radiographs were further analyzed and uploaded on a platform for annotation; bone, crown and root in each radiograph were labeled separately. 1300 radiograph were used for furcation involvement detection, furcations on each radiograph were labeled. Finally, these radiographs were further processed using DensNet and ResNet for bone loss identification and furcation identification, and after identification of bone loss, image analysis was done to predict the amount of bone loss in millimeters.
Results: Accuracy for peri apical radiograph segmentation and bone loss measurement is 92%
accuracy for furcation involvement detection in upper molars is 50%
accuracy for furcation involvement detection in lower molars is 90%
Conclusions: the application of artificial intelligence in detecting and measuring bone loss in periapical radiographs, along with furcation involvement identification, represents a significant advancement in dental radiology. This automated approach provides precise and reliable assessments, assisting clinicians in diagnosing and managing periodontal disease more effectively. The integration of AI technology in dental practices holds great potential for improving patient care and outcomes in the field of periodontology.