<![CDATA[It is very important to protect the human oral cavity to avoid various oral problems, one of which is tumors and oral cancer. Cell growth in the oral cavity is divided into benign oral cavity tumors (benign), precancerous lesions, and oral cavity cancer (malignant). Image classification of benign and malignant lesions can help to determine whether cells in the oral cavity are benign or malignant. CNN is a type of neural network that can be used to extract features from an image. In this research, image classification of benign and malignant lesions will be carried out by applying the ResNet50 architecture to the CNN method. The dataset used is the Oral Image Dataset, which has two classes, namely the benign class and the malignant class. Testing is carried out using testing data from each class using the Adam and SGD optimizers. Based on the test results, it can be concluded that ResNet50 can classify images of benign and malignant lesions well using the Adam optimizer with an accuracy value of 94%.]]>
Copyrights © 2023
