Periodontal Disease Detection with Machine Learning Technology From Radiographic Images : An Interdisciplinary Study Of Dentistry and Computer Science
This study aims to evaluate the effectiveness of the Convolutional Neural Network (CNN) model in identifying periodontal disease using dental images. With the applied method, the CNN model was trained using a dataset consisting of 40 dental images and tested on 55 images to evaluate its ability to...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Universitas Andalas
2024-10-01
|
| Series: | CHIPSET |
| Subjects: | |
| Online Access: | http://chipset.fti.unand.ac.id/index.php/chipset/article/view/280 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This study aims to evaluate the effectiveness of the Convolutional Neural Network (CNN) model in identifying periodontal disease using dental images. With the applied method, the CNN model was trained using a dataset consisting of 40 dental images and tested on 55 images to evaluate its ability to classify the images as healthy or periodontal. The evaluation results showed that the CNN model achieved an overall accuracy of 91.16%. The model precision for healthy images reached 92.39%, while the precision for unhealthy images was 91.05%. Recall sensitivity for healthy images is 91.16%, and for F1-Score images is 91.07%. The data shows that the model has better performance in identifying healthy images compared to periodontal images. To improve the performance of the model, data augmentation techniques such as rotation, flipping, and scaling were applied, which gave a slight improvement to the results. However, the limited size of the dataset seems to be an obstacle in achieving higher accuracy. Therefore, this study recommends expanding the dataset size and applying more complex model architectures or transfer learning techniques to improve detection performance. The conclusion of this study shows that CNN models have potential for periodontal disease detection, but need further development to improve accuracy and reliability. This research contributes to the development of medical detection technology and opens a path for further research in improving periodontal disease detection systems using CNN technology.
|
|---|---|
| ISSN: | 2722-4422 |