Research Problem. This paper specifically aims at providing
suitable features and classification algorithms to decide
which images should be considered as UML CDs and which
images should be left out. The classifier operates by extracting
relevant information about the image and processing that
information with a machine learner. The classifier is expected
to have ability of inclusion of UML CD images and
exclusion of non-UML CD images. Among these tasks,
eliminating non-UML CD images has greater value than
including UML CD images.