Results & Discussion : Identification of dental implants using deep learning (1)
Results
At least 240 instances of each implant system were detected in the panoramic radiographic images: the most common type was MK III/IIIG (1919 instances) and the least common was Genesio (240 instances; Fig. 2). The number of implants detected correctly (True Positive: TP), and those detected as other systems (false positive: FP) are shown in Fig. 3. The number of both TP and FP were the largest in MK III /IIIG and the smallest in Genesio. The TP ratios ranged from 0.50 to 0.82; the highest value was obtained for MK III/IIIG, and the lowest was obtained for Genesio. The values of MK IV/SG and BL were the same (Fig. 4). In Genesio, half of them could not be detected correctly. The APs of each implant were as follows: MK III/IIIG: 0.85, MK IV/SG: 0.78, BL: 0.69, and Genesio: 0.51 (Fig. 5). MK III/IIIG and MK IV/SG could be detected with high accuracy. The mAP and mIoU of this identification system were 0.71 and 0.72, respectively.
Discussion
There are some problems with implants that cannot be solved in general clinics. In these problems, an unknown implant system will make the problems worse. Therefore, the identification of an implant system is necessary for both dentists and patients, and an automated identification system that is not dependent on the dentist’s expertise is needed. Considering these issues, an AI-based approach seems to be a potentially suitable solution, and this study was conducted to focus on developing an automated identification system of implants from panoramic radiographic images using object detection. There are already two methods for implant identification. In the first, dental radiographic images of many implant systems have been uploaded to a website and dentists are able to check them to find an image that matches the patient’s implant image.
Serial posts:
- Identification of dental implants using deep learning — pilot study
- Background : Identification of dental implants using deep learning
- Method : Identification of dental implants using deep learning
- Results & Discussion : Identification of dental implants using deep learning (1)
- Results & Discussion : Identification of dental implants using deep learning (2)
- Results & Discussion : Identification of dental implants using deep learning (3)
- Conclusion : Identification of dental implants using deep learning
- Figure 1: Sample image for calculating IoU (MK III implant)
- Figure 2. Total number of objects of each implant systemin all images.
- Figure 3. The total number of implant systems detected correctly (TPs)
- Figure 4. Ratio of implant systems detected correctly to all detected systems
- Figure 5. Average precision (AP) of each implant system in all images
- Figure 6. Sample images of misdetected implants