So if I understand it correctly, the algorithm replaced color intensity with a different RGB color to highlight the intensity of that pixel by swapping the color contained in it to a different hue, and goes on to change the intensity of that hue to create the banding in a sort of "false color" enhancement? And then it goes on to evaluate the entire image, every pixel in the image to create a false color image that shows intensity variance in the overall image?
It seems to increase contrast by removing steps in the color depth to create a more abrupt transition between colors, like reducing 32 bit to 16 bit, or 16 bit to 8 bit, and that is what is causing the "banding" effect.
So it appears the banding represents different intensities of light rather than different wavelengths, and that would be of no value in identifying various substances or elements.
I downloaded the "black" image to local and ran it through my own stuff, mostly just a contrast enhancement, and read the words "This is what see" in the upper left quadrant of the image:
Before:
After:
ETA: If there is a "pulse" every 12 ms, then that is 83 1/3 pulses per second, so you would need a frame rate of around 100 fps to catch them all, like in a high-speed camera, one of those they use to catch insect or hummingbird wing beats or bullets in flight. Using a camera like that, and stepping through the enhanced frames, you should be able to detect pulses increasing and decreasing.
.
It seems to increase contrast by removing steps in the color depth to create a more abrupt transition between colors, like reducing 32 bit to 16 bit, or 16 bit to 8 bit, and that is what is causing the "banding" effect.
So it appears the banding represents different intensities of light rather than different wavelengths, and that would be of no value in identifying various substances or elements.
I downloaded the "black" image to local and ran it through my own stuff, mostly just a contrast enhancement, and read the words "This is what see" in the upper left quadrant of the image:
Before:
After:
ETA: If there is a "pulse" every 12 ms, then that is 83 1/3 pulses per second, so you would need a frame rate of around 100 fps to catch them all, like in a high-speed camera, one of those they use to catch insect or hummingbird wing beats or bullets in flight. Using a camera like that, and stepping through the enhanced frames, you should be able to detect pulses increasing and decreasing.
.