Disadvantages Of Image Segmentation

Using segmentation in medical images is a very important task for detecting [16] the abnormalities and tracking progress of diseases and surgery planning. Segmentation must not allow regions of the image to overlap. The main aim of medical image segmentation is to study anatomical structure, identify regions of interest, measures tissue volume to measure tumor growth. The main goal of literature survey is to present various image segmentation methods, to highlight the advantages and disadvantages. Thresholding[17] one of the oldest method used for image segmentation. It is based on the gray level intensity value of pixels. Histogram of an image consists of valley and peaks where each peak represents the one region and the valley between the peaks represents the threshold value. On the basis of thresholding value, there are two types of threshold values such as global and local thresholding [18].The main drawback is that only two classes are generated

This approach is based on spatial correlation of pixel therefore segmented result is better than thresholding. Regions are splitted which do not satisfy a given homogeneity criteria. Splitting and merging can be used together and its performance depends on the selected homogeneity criterion. The main drawback is seed is manually selected. Region growing can be so sensitive to the noises that it may cause extracted region to have holes or even is disconnected. Region growing is not often used because it is not sufficient to segment brain structure accurately and robustly. As compared to edge detection methods, region growing methods are simpler and have strongly immune to noise