tsarjak / simulate-correct-colorblindness Goto Github PK

I am so grateful for your excellent work! However, the program collapses when I tried to correct an image with alpha channel.

Solution:
Alpha channel should be ignored when reading an image.

class Utils:
    """
    Couple of utils for loading the images.
    """
    @staticmethod
    def load_rgb(path):
        img_rgb = np.array(Image.open(path)) / 255
        return img_rgb # Here should be img_rgb[:,:,:3] like in load_lms()

    @staticmethod
    def load_lms(path):
        img_rgb = np.array(Image.open(path)) / 255
        img_lms = np.dot(img_rgb[:,:,:3], Transforms.rgb_to_lms())

        return img_lms

The recoloring process you are using, in order to 'correct' the colors,
uses some 'HSV shifting methos', can you describe which it is using equations/pseudocode or Reference
to a scientific Paper???

Hi,

Is it possible for you to add sources/documentation on how you got those transformation matrices. I am trying to do something similar in JavaScript and I have been looking at sources on how to do it.

One of them is https://ixora.io/projects/colorblindness/color-blindness-simulation-research/ and they have derived a different transformation matrix.

Thank you

Self explanatory

Hello,

It looks like the colorsys is not a pip package but is part of the standard lib instead (python2/3)

• https://docs.python.org/2.7/library/colorsys.html
• https://docs.python.org/3.5/library/colorsys.html

Maybe the line from the README: sudo pip install Pillow colorsys numpy

should be replaced with sudo pip install Pillow numpy ?

Take care, awesome project btw =)

Hello,

It appears that pictures are rotated during the transformation:

I haven't found out what is the cause for it yet ^^

But since it occurs for each transformation:

I guess it involves a function that is called regardless of the transformation.

Take care,

great package. I encountered a problem where it silently failed when I used windows style file paths instead of linux. I was running from Pycharm terminal on windows machine.

The first time I ran it appeared to work but did not produce any images.

(Simulate-Correct-ColorBlindness) C:\Users\conwayjp\Documents\code\Simulate-Correct-ColorBlindness>python recolor.py -input "C:\Users\conwayjp\test.png" -output "C:\Users\conwayjp" -run_all
ReColorLib completed running! Check output images in C:\Users\conwayjp

When it was run with linux style path names it worked just fine.

(Simulate-Correct-ColorBlindness) C:\Users\conwayjp\Documents\code\Simulate-Correct-ColorBlindness>python recolor.py -input "C:/Users/conwayjp/test.png" -output "C:/Users/conwayjp" -run_all
ReColorLib completed running! Check output images in C:/Users/conwayjp

also, might be worth noting the name of the opencv package on conda is py-opencv, for newbies.