Evaluation Of Enhanced Curvelet Transform Based Classifiers For Brain Tumor Detection
Brain tumor is one of the diseases that cause death to people. According to World Health Organization (WHO)the causalities are increasing year by year. In this context, it is essential to have an automated approach to detect brain tumor. Many computer-aided approaches came into existence. In this paper, we proposed two approaches for efficient detection of brain tumors from Magnetic Resonance Imaging (MRI) images. We employed an Enhanced Curvelet Transform (ECT) for representing images. The rationale behind this is that the traditional approaches like Fourier transform, wavelet transform and Ridgelet transform have their limitations. For instance, Fourier transform cannot represent data in temporal domain. Wavelets can overcome this problem but cannot represent images with different angles and images. On the other hand, Ridgelets are good to images with line singularities but they are not efficient at handling images with curves. Curvelet Transform (CT) is found to be a good candidate to solve this problem. In this paper, we enhanced CT for better results. Two methods are defined based on ECT for brain tumor detection (BTD). First, a traditional approach is used and then Artificial Neural Network (ANN) based approach is implemented. Both are based on ECT and they are evaluated and the results are provided. The implementation is made using MATLAB. Results revealed that the proposed methods are good for DTD and can be compared with the state of the art.