Identifikasi Penyakit Kanker Mulut (Carsinoma Cell Squamosa) Berdasarkan Kelainan Sel Jaringan dengan Menggunakan Metode Convolutional Neural Network

Abstract

In this study, image lesions were identified as an early prevention of oral cancer (squamous cell carcinoma) using the Convolutional Neural Network algorithm. The initial step that must be done is to analyze the data that will be used for training, then analyze several stages of image processing used, such as image processing, feature extraction with the Gray Level Co-occurance Matrix feature extraction method and the implementation of the algorithm that will be used. The amount of data processed is 425 images of oral lesions which have a size of 250×250 pixels in *.jpg format. After testing the identification of the oral lesions dataset, the lesion image training was carried out to obtain the GLCM feature extraction value. The experimental results show that the application can read the pixel value of the training image with a maximum network parameter of error of 0.01 and a learning rate of 0.05. The application can identify normal lesion images or Squamous Cell Carcinoma lesions with the best results on Model-4 and Model-5 with an accuracy value of 90%.