Color image analysis via Racah moments

In this research, Racah moments color image descriptors are investigated. Channels separation for color images were involved in RGB color space, and Racah moments from each channel were computed individually. In experiments, we have performed the color image reconstructions from different color channels and various orders of Racah moments. Our experimental results have demonstrated that, by applying Racah moments, color images can be completely reconstructed without any information loss. We have also discovered that the individual set of Racah moments will uniquely describe the color image features, while the lower and higher orders of Racah moments will preserve more of features from the bottom right and top left regions of a color image, respectively.