Handling RGBA images

vision
1

When we have alpha channel present in images, reading the data with ImageFolder converts it to RGB. I’m a Pytorch noobie. How would one write custom Datalodaer/Dataset to read images with 4 channels?

2

what is your image file format?

3

Images are PNG with 32 bit format, 8 bits for each of the four channel.

4

ImageFolder uses the pil_loader by default, if accimage is not the used backend and if no other loader was specified.
This loader converts the images to RGB as seen here.
To use the RGBA channels, you could write a custom loader (basically just copy-paste the pil_loader and remove the convert call or convert to RGBA) and pass it as the loader argument to ImageFolder.

5

This worked. Thanks!

6

Uploading my solution:

from PIL import Image
from torchvision import datasets

def custom_loader(path):
    with open(path, 'rb') as f:
        img = Image.open(f)
        return img.convert('RGBA')
image_datasets = {x: datasets.ImageFolder(os.path.join(data_dir, x),data_transforms[x], 
                  loader=custom_loader)
                  for x in ['train', 'val']}