Brain tumor classification using SVM based AlexNet


R. Anita Jasmine, P. Arockia Jansi Rani
Dept of Computer Science and Engineering, Manonmaniam Sundaranar University,Tirunelveli, Tamilnadu, INDIA.


Advanced MRI techniques is one of the best proven technique in tumor analysis and visualization. More than a decade, several machine learning techniques have been deployed for tumor segmentation, feature extraction and classification. The outcome ofBrain Tumor classification plays a critical role in diagnosis and further treatment. The efficacy of the conventional machine learning algorithms depends on the feature extraction.Spotting the suitable feature extraction technique is a tedious task in Machine Learning. At present, Deep learning networks are highly capable of extracting prominent features automatically from images for classification. In this paper, the pre-trained CNNs is used to classify three types of brain tumors from the bench mark dataset, T1 weighted contrast enhanced MRI. The pre-trained network was experimented using the segmented brain tumor image. Experimental results show that the pretrained network achieves 95% for tumor classification.