A Novel Periodontal Disease Grade Classification Methodology using Convolutional Neural Network

Main Article Content

Saptadeepa Kalita, Ram Chandra Singh, Ali Imam Abidi, Hemant Sawhney

Abstract

With the advancement of artificial intelligence, the demand of automated assistance in the domain of medical imaging has become important to reduce time and inaccuracy in physical examination. Modern lifestyle choices are resulting in a variety of dental diseases which led to multiple researches challenges which are carried out for pre-emptive detection in order to deal with these diseases within time. Periodontal diseases (PD) is a type of dental disorder which is rapidly increasing and being the major cause of early teeth loss. The machine learning based convolutional neural network (CNN) model is carried out in detecting grade wise classification of the periodontal disease. A dataset containing 350 dental images in Radio Visio Graphy (RVG) format belonging to an age group of 18-75 years is used for both training as well as testing. This method has successfully detected mild periodontitis, moderate periodontitis and severe periodontitis by achieving a satisfactory accuracy of as high as 94% with minimum loss, precision value, recall and F1score of 0.41, 0.93-0.95 and 0.91-0.94 respectively for all the three classes of Periodontitis.

Article Details

How to Cite
Saptadeepa Kalita, et al. (2023). A Novel Periodontal Disease Grade Classification Methodology using Convolutional Neural Network. International Journal on Recent and Innovation Trends in Computing and Communication, 11(10), 1948–1953. https://doi.org/10.17762/ijritcc.v11i10.8806
Section
Articles
Author Biography

Saptadeepa Kalita, Ram Chandra Singh, Ali Imam Abidi, Hemant Sawhney

Saptadeepa Kalita1, Ram Chandra Singh2, Ali Imam Abidi 1, Hemant Sawhney 3

1Department of Computer Science and Engineering, Sharda University, Greater Noida

2Department of Physics, Sharda University, Greater Noida

3Department of Oral Medicine and Radiology, School of Dental Science, Sharda University, Greater Noida

Corresponding author: ali.abidi@sharda.ac.in