Smartphone-Based Method for Detecting Periodontal Disease

Abstract

In this paper, we propose a novel periodontal disease detection method using smartphones, image processing, and machine learning techniques. Periodontal disease is an inflammatory disease known to be the main cause of tooth loss. Here, a CIELAB color space is adopted for feature extraction and the support vector machine (SVM) is applied for distinguishing healthy gum from diseased gum. A gadget is designed to block ambient light and eliminate refraction effect as well. We recruited 30 subjects consisting of 15 gum-diseased and 15 healthy subjects. Experimental results show that our proposed method detects periodontal infection with 94.3% accuracy, 92.6% sensitivity, and 93% specificity, respectively.