Main motivation behind this investigation is to determine how and to what extent Gray Component Removal (GCR) influence a color image. Specifically how GCR influence image in spatial domain, as well as in Fourier and Cosine domains. This is important to determine in order to develop a data hiding technique that can be used in GCR communication channel. To evaluate the influence of GCR on color image a large image dataset is first converted from RGB to CMYK color space with very aggressive GCR, then the original and the converted images are compared in Fourier and Cosine domains.
The results show that the GCR communication channel has nonlinear impact on the image that depends exclusively on the content of the image. The impact in spatial domain is too severe for watermarking techniques in spatial domain. In contrast, the extent of the impact in frequency domain is low enough so GCR can be used for data hiding techniques based on both, DFT and DCT transforms.