A new algorithm for automatic extraction of interesting objects is proposed in this paper. The proposed algorithm can be summarized in two steps. First, segmentation of color image discriminating interesting objects and backgrounds is performed. According to the research stating, 'humans perceive things by contracting them into three to four essential colors,' a color image is segmented into three regions utilizing k-mean algorithm, followed by the merger of the regions performed when their similarities exceeds the critical value that is drawn from the calculation of the histogram similarity. Second, identifying an interesting object out of the segmented image, generated upon the image composition theory, is performed. To have a good picture, it is important to adjust positions of interesting objects as the picture composition theory. Extracting objects is a retro-deduction process using a weighted mask based on the triangular composition of picture. To show merits of the proposed method, experiments are conducted over 400 images in comparison with recently proposed k-means connectivity constraint and graph-based image segmentation methods. © 2008 Elsevier Ltd. All rights reserved.
Mendeley saves you time finding and organizing research
Choose a citation style from the tabs below