IMPLEMENTASI MODEL EFFICIENTNETV2B0 UNTUK KLASIFIKASI TUMOR OTAK PADA CITRA MRI

Accurate and timely brain tumor classification is essential for early diagnosis and appropriate medical treatment. This study aims to classify brain tumor types in MRI images using the EfficientNetV2B0 architecture, enhanced with data augmentation techniques and CLAHE (Contrast Limited Adaptive Histogram Equalization) preprocessing. The model classifies MRI images into four categories: no tumor, glioma, meningioma, and pituitary tumor. Three training scenarios were conducted: (1) without augmentation, (2) with augmentation, and (3) combining augmentation and CLAHE. Performance evaluation showed a progressive improvement in accuracy, from 95% in the first scenario to 96% in the second, and reaching 97% in the third. These results were also supported by high and consistent precision, recall, and F1-score values. The findings demonstrate that the combination of EfficientNetV2B0, data augmentation, and CLAHE significantly enhances the accuracy and robustness of brain tumor classification in MRI images. This approach is expected to contribute to the development of decision support systems in medical radiology.