KLASIFIKASI JENIS-JENIS TUMOR OTAK MENGGUNAKAN MODEL ARSITEKTUR INCEPTION-V3

The brain is in important part of the body in humans in which there are cell cells that are different from their abilities. Cells that can collectively from tissue and unusual cell doublings from tumors that would become cancerous or non-cancerous cells. The cells are classified into 4 grade classes. Grades I dan II can be represented as lower grade with the designation Benign tumors and Grade III and IV as Malignant Tumors with high tumor cell levels. Medical imaging such as Magnetic Resonance Image (MRI) is a diagnostic method to treat a patient’s condition by manually reviewing images and time-consuming and also prone to errors. In overcoming this problem, a deep learning system is needed to ease the workload of MRI classification and diagnosis. This study aims to classify using the Convolution Neural Network (CNN) method using the Inception-V3 architecture by adding augmentation and hyperparameter tuning to get the best performance. The results showed that the Inception-V3 model by applying augmentation and hyperparameter tuning can improve performance better than before with an accuracy of 89%. The application of augmentation and hyperparameter techniques is proven in increasing accuracy and reducing the occurrence overfiting in the proposed model