A seeded region growing algorithm for spot detection in medical image segmentation

The novel and robust seeded region grow segmentation technique is implemented for spot detection in Computerized Tomography scan image. Segmentation of scanned medical image is quite difficult task due to presence of speckle noise. The grey scale image is segmented into regions, with each region having closest seed correlation. The algorithm is rapid, free of tuning parameters and pixel order dependent. Experimental performance evaluation through manual seed selection demonstrates accurate and efficient segmentation results for chest scan image.