Thanks, Kali! The algorithm varies from case to case, but I can describe the approach broadly by comparing it to the more usual method.
STANDARD COLORING
The program has a way of determining directly the amount the red, green, and blue for each pixel of the output image. That is, based on results from iterative calculations for a particular pixel, it computes
red = . . . on a scale of 0 to 255
green = . . . on a scale of 0 to 255
blue = . . . on a scale of 0 to 255
IMAGE-BASED COLORING
Assume the input bitmap image has NROW rows and NCOL columns. The software makes use of similar formulas
row_number = . . . on a scale of 1 to NROW
col_number = . . . on a scale of 1 to NCOL
Iterative fractal calculations are performed for a particular output pixel. Based on the results, the program then evaluates these two formulas. To color the pixel of the
output image, it just grabs the red/green/blue values from the pixel of the
input image located at the computed row_number and col_number.
Clear as mud, I'm sure.
Regards,
Ken