Adaptive ۱-D polynomial coding to compress color image with C۴۲۱

Color images, in spite of it, are quick and simple to understand (interpret) by our minds, but digitally suffer from huge byte consumptions that directly affect storage (space) and/or transmission. Color data compression systems are vital to control the enormous size of color data by losing insignificant color data that implies redundancies and manipulating significant color information. This research is concerned with reducing or packing the storage of color visual natural images using color transformation of YC_bC_r base, based on a hybrid spatial-frequency compression technique. The essential part of the source band that proposed compression system constitutes of exploiting the adaptive lossy ۱-D polynomial coding for coefficients and Minimize Matrix Size Algorithm (MMSA) for residual. On the other hand, for the non-source bands, the improved hierarchal scalar quantization method called double scalar uniform quantization scheme (DSUQS) is exploited along C۴۲۱. For testing the performance of the suggested system, four natural color images from two datasets were adopted, the first called (Miscellaneous) and the second called (Kodak) of square sizes either of (۵۱۲\times ۵۱۲) pixels. The experimental results showed the superiority of YC_bC_r compared to the well-known standard joint photographic expert group (JPEG) that achieved higher performances where CR between (۲۱-۲۷) and PSNR between (۳۶-۳۸) dB, where CR of JPEG between (۱۶-۲۱) and PSNR (۳۰-۳۳) dB that implicitly means that the proposed system of color transformation base improved on average compared to JPEG with better-perceived quality for the compressed image.

Keywords:

Color Image , YCbCr , MMSA , C۴۲۱ , DSUQS , Miscellaneous , Kodak