In this section, we introduce some of the methods used
in our comparative study; the details can be found in the
literature. As mentioned above, the methods leveraged here
belong to the pixel-based category of CADs, in which each
pixel is represented as a separate feature vector and is classified
individually. This approach should be contrasted to the regionbased approach which forms regions, e.g., by segmentation,
then a feature vector is formed for each region.