Analysis of Brain Disorder Neuroimages by Feature Extraction

Medical images provide significant information for diagnosis of the disease. For medical practitioners always, it was a challenging and time consuming task to analyze the brain tumor. To overcome the same, we proposed a methodology to analyze the brain tumor. Studying the brain with all verticals is challenging, as focusing on the featured execration of information makes the work easy to diagnose brain disorders. Several researchers have implemented methods for analyzing brain disorders; most methods have certain limitations. After performing the literature review, we proposed our unique methodology in which types of neuroimages and classifiers are implemented. In this research work, neuroimages features are extracted from the histogram and morphological based approach. Performance of the research work concluded based on statistical parameters of FCM and watershed classifier. Experiments are performed on brain disorders with MRI and CT modalities. The effectiveness of the proposed method has been tested on MRI and CT image datasets. The algorithm appraises two datasets; MRISIT for magnetic resonance imaging and unpaired MR-CT brain dataset for unsupervised image translation for computed tomography, from which features are extracted by formulated calculation of mean, standard deviation, correlation coefficient, homogeneity, contrast, entropy, energy, and performance measured by sensitivity, specificity, precision, and accuracy. Next, covers a review of existing state-of-the-art work done and discussed the future scope of the proposed system. Further, the quantitative analysis was done in terms of accuracy, which was achieved after algorithm implementation for FCM 86.84% and watershed 83.10%.