Simulasi Deteksi Tonsilitis Mengunakan Pengolahan Citra Digital Berdasarkan Warna dan Luasan pada Tonsil

Simulasi Deteksi Tonsilitis Mengunakan Pengolahan Citra Digital Berdasarkan Warna dan Luasan pada Tonsil

Author		: SANG MADE LANANG PRASETYA; I NYOMAN APRAZ RAMATYANA
Published on	: Jurnal Nasional Teknik Elektro dan Teknologi Informasi, UGM

Abstract

Tonsillitis or known as tonsils is a medical condition characterized by inflammation of the tonsils, causing sore throat, difficulty swallowing, fever, and in certain cases can lead to heart attack or pneumonia. Doctors diagnose tonsillitis in a visual way, see tonsil inflammation and assess subjectively. In this study designed a tool to calculate the area of inflamed areas that can be used to help doctors diagnose tonsillitis. Tonsils image processed on the red layer to quantify the extent of tonsils. Furthermore, the red area is calculated as area of inflammation. In next stage, find the feature extraction using histogram analysis to find the distribution of image intensity levels. The results were classified using k-Nearest Neighbor (k-NN). From 64 datas which consists of 32 normal and 32 tonsillitis, a system can reached 90,625% accuracy rate. This value is achieved at the cityblock distance measurement and k = 1.