ABSTRACT
Image segmentation in computer vision applications can generally be described as the process of di›erentiating an object of interest in an image or a volume from its background. Computer vision researchers generally agree that segmentation is one of the most challenging and important tasks in the œeld of computer vision. Historically, many computer vision problems have been roughly broken down into a three-step process: (1) segmentation, (2) feature extraction and analysis, and (3) recognition (or classi-œcation). Segmentation, being the œrst step in the process in this serialized approach, has direct impact on all tasks that follow. If an object is not accurately segmented from its background, the subsequent tasks of feature extraction and classiœcation may become meaningless or, even worse, inaccurate. Hence, the segmentation task is critical to the success of the entire analysis. To compound matters further, it is o£en the most di—cult task of the three.
