Dental Caries early detection using Convolutional Neural Network for Tele dentistry
- Resource Type
- Conference
- Authors
- Saini, Devesh; Jain, Richa; Thakur, Anita
- Source
- 2021 7th International Conference on Advanced Computing and Communication Systems (ICACCS) Advanced Computing and Communication Systems (ICACCS), 2021 7th International Conference on. 1:958-963 Mar, 2021
- Subject
- Communication, Networking and Broadcast Technologies
Components, Circuits, Devices and Systems
Computing and Processing
General Topics for Engineers
Power, Energy and Industry Applications
Robotics and Control Systems
Signal Processing and Analysis
Training
Deep learning
Visualization
Pain
Neural networks
Teeth
Medical services
Dental caries
Artificial neural network (ANN)
Soft Computing techniques
Convolution Neural Network (CNN)
- Language
- ISSN
- 2575-7288
Dental caries is progressive bacterial infection which is the main cause of tooth lose. This occurs due to poor dental hygiene which also lead to various dental diseases. If caries can be diagnosis in early stage using tele-dentistry system it will great help for child oral care. Because severe form of caries produces infection and pain which may be result of tooth extraction. So early detection and diagnosis of these caries are the major concern for the researchers. Soft computing techniques are prominently used in dentistry which makes diagnosis easy and reduces the diagnosis time. This paper aims to detect dental caries using digital color image at an early stage so that the treatment can be performed easily and effectively. This classification also suitable for tele dentistry as tele informatic oral health care system for this proposed work we implemented the deep learning model is convolution neural network (CNN). We have trained deep learning different models of CNN; they are Visual Geometry Group (Vgg16 & Vgg19) Inception-V3 and Resnet50. Training, validation, and testing has been performed on binary dataset with caries and without caries images. The classification accuracy is achieved using Vgg16, Vgg19 Inception v3 and Resnet50 models and the highest accuracy is achieved among them is by Inception v3 with the training accuracy of 99.89% and validation accuracy of 98.95 with minimum loss compared to Vgg16 CNN models.