<![CDATA[ Nowadays medical science has developed rapidly, diagnostic and treatment techniques have provided life expectancy for patients. One way of examining brain tumor sufferers is radiological examination that needs to be done, including MRI with contrast. MRI brain images are useful for seeing tumors in the initial steps of diagnosis and are very good for classification, erosions / destruction lesions of the skull. Smoothing image processing, segmentation with otsu method and feature extraction are carried out to facilitate the training and testing process. This study, will apply texture analysis with the parameters contrast, correlation, energy, homogenity to distinguish the texture of brain tumor images and normal so as to produce a standard gold value based on existing texture characteristics. Training and testing of texture features using backpropagation method of artificial neural networks with variations in learning rate values so that it is expected to obtain a classification of the image conditions of patients with brain tumors. The data used are 29 brain images that produce classification accuracy of 96.55%.Keywords : MRI images, brain tumors, textur, backprogation ]]>
Copyrights © 2018
