Klasifikasi Jenis Penyakit Luka Bakar Menggunakan Metode Faster Region Convolutional Neural Network (Faster R-CNN)

Abstract

Burns are one type of injury to the skin caused by heat either from fire, exposure to chemicals, electric shock to solar radiation. Burns are also one of the most common types of injuries that occur accidentally. Burns are one type of external injury that often occurs in humans. There are several types of burns with different treatments in each type of wound. Due to treatment, especially the wrong first handler in treating one type of burn, many side effects occur after the wound heals such as skin that has the potential for infection, unnatural scars and makes the wound heal longer. This research resulted in a system that can classify burns based on the color and shape of the wound. The types of burns are divided into three levels, namely Degree I, Degree II and Degree III Burns. Grade I Burns in this study were characterized by striped skin color to reddish skin color, Grade II Burns were characterized by reddish to white, more sculptural skin, and Grade III Burns were characterized by black skin color with a rougher texture and a larger area of damage. The data used in this study amounted to 1250 data which were then divided into 1000 training data, 125 validation data, and 125 testing data. After testing, this study resulted in an accuracy of 95.20%.