❓ read_image gives differently shaped arrays based on as_tensor argument.

[JeGa]

Your Question

Is it intended, that read_image returns different shapes (order of c, h, w), depending on the as_tensor parameter?

print(read_image(image_file_path, as_tensor=True).shape)
# torch.Size([3, 490, 326])
print(read_image(image_file_path, as_tensor=False).shape)
# (490, 326, 3)

Forum Check

• I have searched the discussions forum for an answer to my question.

[github-actions]

Thanks for submitting this issue! It has been added to our triage queue. A maintainer will review it shortly.

[rajeshgangireddy]

Hi,
Yes, this is intentional. It follows the standard conventions of each respective library.
Torch Tensor needs to be in [C, H, W] format.

def read_image(path: str | Path, as_tensor: bool = False) -> torch.Tensor | np.ndarray:
    """Read RGB image from disk.

    Args:
        path (str | Path): Path to image file
        as_tensor (bool): If ``True``, return torch.Tensor. Defaults to ``False``

    Returns:
        torch.Tensor | np.ndarray: Image as tensor or array, normalized to [0,1]

    Examples:
        >>> image = read_image("image.jpg")
        >>> type(image)
        <class 'numpy.ndarray'>

        >>> image = read_image("image.jpg", as_tensor=True)
        >>> type(image)
        <class 'torch.Tensor'>
    """
    image = Image.open(path).convert("RGB")
    return to_dtype(to_image(image), torch.float32, scale=True) if as_tensor else np.array(image) / 255.0