Segmentation Stroke Objects based on CT Scan Image using Thresholding Method

Brain image segmentation is one of the most important parts from a clinical diagnostic tool to determine the characteristics of a particular stroke type. Find anatomical contours and the location of the stroke to characterize the type of stroke perfectly in segmentation is very difficult this research proposes an approach to Image segmentation by the process of separating objects from other objects in CT Scan images. CT scan image segmentation uses the thresholding method with the Binarization process. Implementation of the Threshold method is Global binary thresholding, and Otsu thresholding. Preprocessing images to make repairs before segmentation. The dataset is used from Surabaya Hajj General Hospital and public data. The results of this experiment is improved image evaluated using peak signal-to-noise ratio (PSNR) and mean-square error (MSE), the best results were seen in bilateral filtering with a PSNR value of 69% MSE which was the lowest 0.008%. The best stroke object segmentation results using Otsu Thresholding by determining the lower threshold with a High Value of ≤170.