Image Segmentation

Abstract

Image segmentation is one of the major areas of research in image processing and computer vision. It consists in subdividing an image into its constituent parts and is typically used to identify objects or other relevant information in digital images. Segmentation of non-trivial images is one of the most difficult tasks in image processing, due to the great differences between image types. For instance, natural images have very different features from medical images, from the ones obtained for controlling the quality of manufactured goods, or from scanned documents. Image segmentation has several applications such as analyzing the patient’s anatomy from medical images, detecting the layout of an image document, or finding obstacles in a street from a camera installed in a car. In addition, for the same image, different segmentation techniques can be required depending on the regions that the user wants to detect. For instance, if two different anatomical regions have to be segmented in a medical image, then two different segmentation processes will be required.