Detection of Brain Tumor in MRI Image Using SVM Classifier

Abstract

Biomedical imaging plays an important role in diagnosis and early detection of tumor. Brain Tumor is an uncontrolled tissue growth found in any part of the brain. Early stage tumor detection makes the treatment easier. The important imaging techniques are X-ray, computed tomography, magnetic resonance imaging, ultrasound, etc. Magnetic resonance imaging (MRI) is the best technique to detect the tumor portion in the brain. The existing system uses fusion technique of image which is a method of joining complementary information and multi-modality images of the patient. Moreover, the early stages of tumor cannot be detected effectively. In the proposed system, SVM classifier is used in the detection of tumor affected portion. Noise in the acquired image is removed using Gabor filter. The function of Gabor filter is edge detection, feature extraction and noise removal. The morphological function such as dilation and erosion will be applied through the filtered image. Then, the enclosed region will be spitted out separated by the SVM classifier. Using SVM classifier, the early stages of tumor can be easily detected. Also, it is used to segment the tumor portion. This paper also aims at sending notifications to the guardians, about the patient’s condition and the medications to be taken by the patients by means of both mails and messages.