This algorithm provides the means of generatitng camofulage patterns based on the Perlin fractal noise from an input image.
The main.py file contains a scratch of the project in functional programming and it's deprecated, please use the objectmaker.py script.
In order to properly use the algorithm the camo class must be instantiated with the path to the input image.
The input image should be in the ./data/ directory.
In order to extract the colors from the input image the extract_colors(self, n, show=False, save=True, load=False, colspace="HSV") method must be called.
n is the number of colors to extract; the defult color space is HSV however RGB is also supported but yileds worst results.
In order to generate a proper pattern you should use the layer_perlin(self, res=(16, 8), spotting=[], test=False) method.
The res parameters is a tuple of two integers which represent the periods of the noise to be generated along the two dimentional axes.
The spotting parameter is a list used to set the parameters for addittion of spotting to the pattern, it accepts values in the form of [number of colors(int), fraction(float)], if empty no spotting will be performed.
Please note that the only proper method to generate patterns is layer_perlin; al the other methods, including optimization ones yiled suboptimal results and included just for the sake of clarity and compleatness.
Can be runned dowloading the repo, all files are included.
#demo with large forest image
forest = camo("data/forest.jpg")
#first we have to run the color extractor
forest.extract_colors(n=3, colspace="HSV", show=True, load = True)
#now we run the pattern designer asking it to apply spotting the 10%(0.1) of the surface using the two least occurring colors
forest.layer_perlin(spotting=[2, 0.1], res=(4, 4))
forest.show()