Oral Disease Detection using Neural Network

Abstract

Oral diseases such as periodontal, oral cancer and tooth trauma are common non-communicable diseases that affect during the lifetime causing pain, uneasiness and even death. Periodontal is the eleventh most prevalent disease globally and the incidence of oral cancer is estimated around 20 cases per 1,000 people. It's normally caused by poor brushing, hormonal changes and flossing habits that allow plaque, a sticky film of bacteria. The mouth is regarded as a mirror of the complete wellness of the body and till now there is no personal device existing currently to monitor the oral health. An accurate prediction is essential for correct diagnosis and treatment of oral diseases. Dental diseases are mostly diagnosed at the later stage after severe pain occurs in the mouth. By predicting it at early stage, oral diseases which slowly make the roots of the teeth to get weaker can be prevented. This work aims at creating an economical, multimodal, personal oral sensing device that automatically senses and categorizes the data which will assist the clinician in early diagnosis and effective treatment. Our proposed smart electronic device automatically captures valuable parameters like pH, temperature, CO2 and other gases to overcome the challenges in the diagnosis of the oral problem. The captured data is fed to Convolutional Neural Network for classification of oral diseases.