DIAGNOSTIC CHARTING ON PANORAMIC RADIOGRAPHY USING DEEP- LEARNING
ARTIFICIAL INTELLIGENCE SYSTEM
Abstract
Aims of the Study: A radiographic examination is a significant part of
the clinical routine for the diagnosis, management, and follow-up of the
disease. Artificial intelligence in dentistry shows that the deep
learning technique high enough quality and effective to diagnose and
interpret the images in the dental practice. For this purpose, it is
aimed to evaluate diagnostic charting on panoramic radiography using a
deep-learning AI system in this study. Methods: 1084 anonymized dental
panoramic radiographs were labeled for 10 different dental situations
including crown, pontic, root-canal treated tooth, implant,
implant-supported crown, impacted tooth, residual root, filling, caries,
and dental calculus. AI Model (Craniocatch, Eskişehir, Turkey) based on
a deep CNN method was proposed. A Faster R-CNN Inception v2 (COCO) model
implemented with Tensorflow library was used for model development. The
training and validation data sets were used to predict and generate
optimal CNN algorithm weight factors. Results: The proposed artificial
intelligence model has promising results for detecting dental conditions
in panoramic radiographs except for caries and dental calculus. The most
successful F1 Scores were obtained from the implant, crown, and
implant-supported crown as 0,9433, 0,9122, 0,8947, respectively.
Conclusion: Thanks to the improvement of the success rate of AI models
in all areas of dentistry radiology, it is predicted that they will help
physicians especially in panoramic diagnosis and treatment planning, as
well as digital-based student education, especially in this pandemic
period when online training is on our agenda.