An Ensemble Deep Neural Network Approach for Oral Cancer Screening

Nanditha B R,Dinesh M S,Geetha Kiran A,Murali S,Chandrashekar H S

doi:10.3991/ijoe.v17i02.19207

Abstract

<p class="0abstract"><span lang="EN-US">One of the ways to reduce oral cancer mortality rate is diagnosing oral lesions at initial stages to classify them as precancerous or normal lesions. During routine oral examination, oral lesions are normally screened manually. In a low resource setting area where there is lack of medical facilities and also medical expertise, an automated mechanism for oral cancer screening is required. The present work is an attempt towards developing an automated system for diagnosing oral lesions using deep learning techniques. An ensemble deep learning model that combines the benefits of Resnet-50 and VGG-16 has been developed. This model has been trained with an augmented dataset of oral lesion images. The model outperforms other popularly used deep learning models in performing the classification of oral images. An accuracy of 96.2%, 98.14% sensitivity and 94.23% specificity was achieved with the ensemble deep learning model.</span></p>

Highlights

Cancer has become a serious global health disorder that has a high incidence of mortality
The 5-year survival rate of patients diagnosed with oral cancer at an early stage is 82%, and that of patients diagnosed in later stages is 27% in India [6]
An ensemble deep learning model to classify oral lesions into premalignant and normal lesions from digital true color images has been proposed in the present work

Summary

Introduction

Cancer has become a serious global health disorder that has a high incidence of mortality. Nearly 6.5 lakh people are diagnosed with oral cancer and about 3 lakh deaths are reported annually [1]. The USA [3] has 67% more survival rate in the last five years and India [4] has 37%. Detection and treatment at an early stage has resulted in an increase in oral cancer survival rate [5] in India. The 5-year survival rate of patients diagnosed with oral cancer at an early stage is 82%, and that of patients diagnosed in later stages is 27% in India [6]. In a resource limited country like India, where people in rural and remote iJOE ‒ Vol 17, No 02, 2021

Methods

Results

Conclusion