def __call__(self, img: np.ndarray):
"""
Args:
img: numpy.ndarray containing input data. Must be real and in shape [channels, spatial1, spatial2, ...].
Returns:
np.ndarray containing smoothed result.
"""
# add one to transform axis because a batch axis will be added at dimension 0
savgol_filter = SavitzkyGolayFilter(self.window_length, self.order, self.axis + 1, self.mode)
# convert to Tensor and add Batch axis expected by HilbertTransform
input_data = torch.as_tensor(np.ascontiguousarray(img)).unsqueeze(0)
return savgol_filter(input_data).squeeze(0).numpy()
def test_value(self, arguments, image, expected_data, atol, rtol=1e-5):
result = SavitzkyGolayFilter(**arguments)(image.to(device="cuda"))
np.testing.assert_allclose(result.cpu(), expected_data, atol=atol, rtol=rtol)
def test_value(self, arguments, image, expected_data, atol):
result = SavitzkyGolayFilter(**arguments)(image)
np.testing.assert_allclose(result, expected_data, atol=atol)