Automatic Color Images Classification Algorithm
Rushdi Abu Zneit, Amjad Abu Jazar, Belal Ahmad Ayyoub · 2012
Numerous research works about the extraction of objects from images have been published. However only recently the focus has shifted to exploiting low-level features to classify images automatically into semantically meaningful and broad categories. This paper presents a novel automatic color image classification algorithm. Initially the color image is divided into classes; each class is a group of pixels that they have the same color after that the object is extracted from the image. In the recent work, an automatic color images classification algorithm is synthesized and analyzed. The suggested method was running on group of color images to determine the best parameters of suggested algorithm. We run our algorithm with initial value of elaboration coefficient � � 5 and we found that the automatic classification algorithm achieved minimum classification error and minimum running time when � incremented by step d � � 1 and � becomes � 20.