# HSI color space and RGB color space relationship in Steganography

I need to embed text in an image, research papers say HSI produces better mse and psnr therefore are better for steganography. I have RGB values(69,223,90) and its corresponding HSI is (127.223,0.458 ,0.499) I transformed I-plane 0.499 so its between 0 and 255 and got 127 as the Intensity, decreased it by one to 126(embed 0 bit). 126 was converted to the I plane normal range 0.494 and the new RGB values are red=68 green=221 blue=89. It decreased all 3 RGB values and green even more than 1. I dont understand how this is better and makes psnr better? I know the equation but say i decreased/increased the value of the intensity plane by 1 how does this affect RGB?

• Well, think about a vector quantity that is rotated. If your not reading the values in the exact direction of rotation its has smaller amplitude. But were you to analyze the in polar coordinates it does not matter what the rotation is as you are readfing the magnitude and discarding teh angle. So the magnitude is now invariant of the data orientation... HSL is a polar transform. Perhaps is should convert this to an answer but its just my logical reasoning not very deep thinking. – joojaa May 25 '20 at 17:35