CT Scan Based Brain Tumor Recognition and Extraction using Prewitt and Morphological Dilation

A tumor arises when there is an unbounded splitting up of cells that create an abnormal cluster of cells around or within the brain and this cell group may influence the functioning of the brain by destroying healthy cells which is genuinely life-threatening. Brain tumor affects millions of populace, so the early prediction is extremely necessary for patients. There is a must required for an accurate and economical system to predict the tumor; it will be very subsidiary for taking necessary precautions and treatment, so the primary intension of the developed system is automatic identification of the tumor from the brain CT scan image. There are different kinds of medical imaging methods that are competent to produce a visual representation of inside the body, CT scan is one of them which permit look internally but it has been recognized manually by radiologist. Manually lump finding has the possibility of human mistake, due to the akin gray level of an image; radiologists may puzzle, so manually detection is not an easy job. To plummeting this difficulty we develop a new approach that is competent enough to differentiate tumor (malignant or benign) and healthy brain automatically through an image processing approach. Various stages involved in the developed methodology are Pre-Processing, Segmentation, Prewitt Edge Recognition Filter, Morphological operation, Flood filling, Tumor Extraction. Prewitt is a gradient-based edge recognition filter that is beneficial for identifying the area of picture where large amends in intensity happens. It is also subsidiary for segmenting and identification of an object in a scene after that applied dilation operation for filling petty holes by modifying image pixel intensity and generate a suitable image for tumor extraction. Usually, existing approaches may distort the genuine detail that degrades the system accuracy but the experimental outcome showed that better result compared to existing systems in terms of exactness. The System obtained 94.23% accuracy by testing 52 CT scan sample images that have the tumor affected brain as well as healthy brain.