Image proccessing

Before embedding input images in prototypical networks, images are loaded and processed. I do not understand this code ‘x = 1.0 - torch.from_numpy(x)’ in data loading function. I am new in this field. Can you explain its function. Thank you.

x = x.rotate(float(rot))
x = x.resize((28, 28))

shape = 1, x.size[0], x.size[1]
x = np.array(x, np.float32, copy=False)
x = 1.0 - torch.from_numpy(x)
x = x.transpose(0, 1).contiguous().view(shape)

The mentioned line of code subtracts the x array from a constant scalar 1.0.
It’s a bit hard to tell why this is used without seeing any context.