The mechanism of this project is simple.
It gives you the closest image to your reference image by calculating an absolute grayscale distance to the reference.
The difference from a normalized grayscale distance it that pixel count cannot influence the final comparison while the absolute distance can give you significantly different results from what you expect. (the normalized image computation is to be implemented later in the project)
Before calculating the grayscale distance, the images are transformed into a dictionary of 256 sets, with the key being the grayscale opacity and the value of the set being the number of occurrences of that grayscale value.
