Authors : Shimal Das and Dibyendu Ghoshal
Abstract: Fractal geometry has been found to be an unbelievable breakthrough in solving the segmentation and compression problems of digital images. In present study, effect of variations of colour schemes has been studied with the help of various types of geometrical topology for the basic segmentation process. The most and popular RGB colour space performance has been compared with HSI colour space along with quad tree and isosceles geometrical segmentation. It has been found that hue-situation intensity scheme is superior to RGB schemes for the measurement of PSNR. The execution time for isosceles triangle based geometrical segmentation yield lower processing time. The major efficiency of the study is Multifoods compression ratio when HSI scheme has been adopted. Both types of segmentation viz. Quad tree and isosceles triangle approach have produced nearly four times increment in compression ratio. During the transformation all the channels, e.g., RGB and HSI have been treated separately. The results have been compared with those available from recent literature and it has been found that the proposed schemes of sequence of studies are superior to the earlier ones.
Shimal Das and Dibyendu Ghoshal, 2016. A Comparative Study on Color Image Compression Based on Fractal Geometrical Topologies for Partition Iteration. Asian Journal of Information Technology, 15: 4719-4731.