Comparative Analysis on Brain Tumor Classification using Deep Learning Models
Madona B. Sahaai, G. R. Jothilakshmi, R Selva Kumar, S Praveen Kumar
Abstract
The categorization of brain tumors is crucial for accurate medical analysis as well as healing. Convolutional Neural Network plays an essential role in diagnosing disease in the domain of deep learning algorithms which is extremely pertinent for visual imaging analysis. Initially, the features are extracted from brain MRI images via CNN. In this work, we applied four deep learning based network models such as Dense Net 201, VGG-19, Xception, Inception v3 for brain tumor classification. Comparison had done on four deep learning models based on accuracy to estimate which model generates good results. Finally, experimental outcomes illustrate that DenseNet201 outperforms better accuracy as 91.94% in diagnosing brain tumor and also classification. Moreover, metrics such as precision, recall and F1 score were evaluated to predict the overall performance of the model.