ABSTRACT
Image segmentation is generally used to separate an image into different parts, or to extract significant information from the image. The purpose of thresholding is to mark pixels that belong to foreground pixels with the same intensity and background pixels with different intensities. The histogram method presumes that there is some average value for the foreground or object pixels and background, but the reality is that the real pixel values have some deviation around the average values. The Roberts edge detection technique is used to highlight high spatial frequency regions of the image, which corresponds to edges. Region-based segmentation is used to split or merge regions in the image based on some similar or common image properties such as intensity values of the region, texture, or pattern of the region. This can be divided into two main types: region growing or region merging and region splitting.
