Describe the bug
When creating a test image with data/synthetic/create_test_image_2d, it looks like the dimensions of the image are defined like (width, height) instead (height, width).

This way, it can be a little confusing when creating test images. For example, in NumpyImageTestCase2D, we pass the dimensions as create_test_image_2d(self.im_shape[0], self.im_shape[1]... in

and inside create_test_image_2d, self.im_shape[0] is assign as width and self.im_shape[1] as height, because of the ordering that the args are defined in

But when we create a test image using NumpyImageTestCase3D, we still pass the dimensions as create_test_image_3d(self.im_shape[0], self.im_shape[1], self.im_shape[2], ... in

but this time, self.im_shape[0] is assign as height and self.im_shape[1] as width because of the Args ordering in

Additional context
It can make a little confusing to interpret tests like these:

or when creating tests to check the shape of the output of the networks.