In the current work, seven different color spaces were investigated including sRGB, YCbCr, YCoCg-R, RDgDb, CIELAB, IPT, and CAM02-UCS by incorporating into JPEG1992, JPEG2000 and new JPEG. The prime objective of this work was to compare the performance of CSI-JPEG with that of JPEG1992 and JPEG2000. Psychophysical experiments were also performed to examine visual appearance of the compressed and decompressed images. Compression performance was examined for different color spaces by using measures like SNR, PSNR, MSE, CIEDELAB and CIEDE2000. The results showed that CSI-JPEG algorithm provided about 20-30% more compression rate compared to JPEG1992 algorithm for same color difference. Moreover, CAM02-UCS was found to perform best in terms of compression rate and image quality for all three algorithms. The CAM02-UCS was found significantly better compared with linear color spaces but CIELAB, CIELUV, and IPT were insignificantly better. This also implies that it is visually more uniform than CIE uniform color spaces. Results also showed that performance difference of CSI-JPEG and JPEG2000 was not significant but both performed significantly better compared with JPEG1992.
Muhammad Safdar, M. Ronnier Luo, Xiaoyu Liu, "Performance Evaluation of JPEG, JPEG2000 and New CSI-JPEG Algorithms by Incorporating Different Color Spaces" in Proc. IS&T 23rd Color and Imaging Conf., 2015, pp 219 - 223, https://doi.org/10.2352/CIC.2015.23.1.art00039