UnalignedDataset image_A = Image.open(A_path).convert('L')

KupynOrest / DeblurGAN

Image Deblurring using Generative Adversarial Networks

Other

2.5k stars 516 forks source link

UnalignedDataset image_A = Image.open(A_path).convert('L') #47

Open LinChen-Cherry opened 6 years ago

LinChen-Cherry commented 6 years ago

When use unaligeneddataset, image A is loaded as 'L' mode, while image B is loaded as 'RGB' mode:

    A_img = Image.open(A_path).convert('L')
    B_img = Image.open(B_path).convert('RGB')

Is it a bug?

ruiwen0615 commented 6 years ago

I have the same question. When I use unaligned mode to train, it reports error that input images are not 3 channels. I think the reason might be this conversion?

KupynOrest commented 6 years ago

thanks for catching this bug, I would add a quick fix, you should load the first image in 'RGB' too