In order to model the slow-smoke color, luminance channel is subtracted from chrominance channel and linear interpolation is performed to fit the range of values to [0,255].
Lab color space is used for luminance and chrominance channels separation.
In this space, L denotes luminance while a denotes red-yellow and b represents blue-yellow chrominance channels. A split operation is performed to separate luminance and chrominance channels.
Therefore, we transform the image into YCbCr, take the luminance channel as a grayscale image for the 2D-DCT transformation and computation of the FM and IM values, and take the chrominance channel
to find the CM values.
Since the human visual system is more sensitive to the luminance channel, we use the superresolution reconstruction method based on dictionary learning proposed in this study to reconstruct the images in the luminance channel, and the images in the Cb and Cr chrominance channels
are simply magnified by the bicubic interpolation method.
The processed L channel is then combined with the chrominance channels
('a' and 'b') and converted back to RGB color space.
L from CIE LAB, V from HSV or Y from YIQ, as the chrominance channels
usually do not contain the data useful for further binarization and recognition.
CIE space is suitable for this method because it offers a clear separation of luminance and chrominance channels, allowing for a better control over contrast differences.
H calculates grayscale intensity modified in Fourier domain, E, using Fourier transform values applied over luminance and each of two chrominance channels. H is calculated on each frequency using the next relation:
where [theta] controls chromatic contrast level of influence in the final result, and [PHI] is a coefficient that determines the relative contribution of chrominance channels a and b.