Differentiable Color Conversion for Pytorch

Requirement

1. pytorch
2. numpy

Usage

1. import color.py

2. embededing the conversion class in to you code.

   class ColorModel(nn.Module): ''' input should be B, C, H, W and the value should between [0, 1] output is B, C, H, W and the RGB is between [0, 1] '''

    def __init__(self):
        super(ColorModel, self).__init__()
        self.A = Autoencoder()
        self.rgb = lab2rgb()
        self.lab = rgb2lab()
        
   
    def forward(self,x):# input rgb [0,1]
   
        lab_x = self.lab(x)
        a = self.A(lab_x)
   

TODO

1. often-used color space conversion
2. usage demo
3. Chainer Version
4. Tensorflow Version