Fuzzy Ann Based Detection and Analysis of Pathological and Healthy Tissues in Flair Magnetic Resonance Images of Brain

A computational technique is proposed for the segmentation, detection and analysis of pathological tissues, healthy tissues and Cerebrospinal fluid (CSF) of brain with the help of FLAIR brain magnetic resonance images. Composite feature vectors are extracted from the blocks of size 4 × 4 pixels of intra-cranial brain image. Then using fuzzy C mean algorithm, clustering of feature vectors and segmentation of images are done for five regions tumor, edema white matter(WM), gray matter(GM) and CSF. Feature vectors comprise of empirically developed higher order wavelet function using Daubechies wavelet transform and statistical functions. Analysis of pathological regions and healthy tissue regions are done to see the complication of the disease. Segmentation result is validated and tested for 22 brain tumor patients. Training of artificial neural network using fuzzy back propagation algorithm is done for the detection of tumor, edema, WM, GM and CSF. The result using the composite feature vector developed is very satisfactory and mean square error is 2.58952 e-013.