def test_structural_similarity_grad(seed):
N = 30
# NOTE: This test is known to randomly fail on some systems (Mac OS X 10.6)
# And when testing tests in parallel. Therefore, we choose a few
# seeds that are known to work.
# The likely cause of this failure is that we are setting a hard
# threshold on the value of the gradient. Often the computed gradient
# is only slightly larger than what was measured.
# X = cp.random.rand(N, N) * 255
# Y = cp.random.rand(N, N) * 255
rnd = np.random.RandomState(seed)
X = cp.asarray(rnd.rand(N, N) * 255)
Y = cp.asarray(rnd.rand(N, N) * 255)
f = structural_similarity(X, Y, data_range=255)
g = structural_similarity(X, Y, data_range=255, gradient=True)
assert f < 0.05
assert g[0] < 0.05
assert cp.all(g[1] < 0.05)
mssim, grad, s = structural_similarity(X,
Y,
data_range=255,
gradient=True,
full=True)
assert cp.all(grad < 0.05)
def test_structural_similarity_patch_range():
N = 51
rstate = cp.random.RandomState(1234)
X = (rstate.rand(N, N) * 255).astype(cp.uint8)
Y = (rstate.rand(N, N) * 255).astype(cp.uint8)
# grlee77: increase value from 0.1 to 0.15
assert structural_similarity(X, Y, win_size=N) < 0.1
assert_equal(structural_similarity(X, X, win_size=N), 1)
def test_mssim_mixed_dtype():
mssim = structural_similarity(cam, cam_noisy)
with expected_warnings(["Inputs have mismatched dtype"]):
mssim_mixed = structural_similarity(cam, cam_noisy.astype(cp.float32))
assert_almost_equal(mssim, mssim_mixed)
# no warning when user supplies data_range
mssim_mixed = structural_similarity(cam,
cam_noisy.astype(cp.float32),
data_range=255)
assert_almost_equal(mssim, mssim_mixed)
def test_structural_similarity_dtype():
N = 30
X = cp.random.rand(N, N)
Y = cp.random.rand(N, N)
S1 = structural_similarity(X, Y)
X = (X * 255).astype(cp.uint8)
Y = (X * 255).astype(cp.uint8)
S2 = structural_similarity(X, Y)
assert S1 < 0.15 # grlee77: increase value from 0.1
assert S2 < 0.15 # grlee77: increase value from 0.1
def test_invalid_input():
# size mismatch
X = cp.zeros((9, 9), dtype=cp.double)
Y = cp.zeros((8, 8), dtype=cp.double)
with pytest.raises(ValueError):
structural_similarity(X, Y)
# win_size exceeds image extent
with pytest.raises(ValueError):
structural_similarity(X, X, win_size=X.shape[0] + 1)
# some kwarg inputs must be non-negative
with pytest.raises(ValueError):
structural_similarity(X, X, K1=-0.1)
with pytest.raises(ValueError):
structural_similarity(X, X, K2=-0.1)
with pytest.raises(ValueError):
structural_similarity(X, X, sigma=-1.0)
def test_structural_similarity_multichannel_chelsea():
# color image example
Xc = cp.asarray(data.chelsea())
sigma = 15.0
Yc = cp.clip(Xc + sigma * cp.random.randn(*Xc.shape), 0, 255)
Yc = Yc.astype(Xc.dtype)
# multichannel result should be mean of the individual channel results
mssim = structural_similarity(Xc, Yc, multichannel=True)
mssim_sep = [
float(structural_similarity(Yc[..., c], Xc[..., c]))
for c in range(Xc.shape[-1])
]
assert_almost_equal(mssim, np.mean(mssim_sep))
# structural_similarity of image with itself should be 1.0
assert_equal(structural_similarity(Xc, Xc, multichannel=True), 1.0)
def test_gaussian_structural_similarity_vs_IPOL():
# Tests vs. imdiff result from the following IPOL article and code:
# https://www.ipol.im/pub/art/2011/g_lmii/
mssim_IPOL = 0.327309966087341
mssim = structural_similarity(
cam, cam_noisy, gaussian_weights=True, use_sample_covariance=False
)
assert_almost_equal(mssim, mssim_IPOL, decimal=3)
def test_structural_similarity_multichannel():
N = 100
X = (cp.random.rand(N, N) * 255).astype(cp.uint8)
Y = (cp.random.rand(N, N) * 255).astype(cp.uint8)
S1 = structural_similarity(X, Y, win_size=3)
# replicate across three channels. should get identical value
Xc = cp.tile(X[..., cp.newaxis], (1, 1, 3))
Yc = cp.tile(Y[..., cp.newaxis], (1, 1, 3))
S2 = structural_similarity(Xc, Yc, multichannel=True, win_size=3)
assert_almost_equal(S1, S2)
# full case should return an image as well
m, S3 = structural_similarity(Xc, Yc, multichannel=True, full=True)
assert_equal(S3.shape, Xc.shape)
# gradient case
m, grad = structural_similarity(Xc, Yc, multichannel=True, gradient=True)
assert_equal(grad.shape, Xc.shape)
# full and gradient case
m, grad, S3 = structural_similarity(Xc,
Yc,
multichannel=True,
full=True,
gradient=True)
assert_equal(grad.shape, Xc.shape)
assert_equal(S3.shape, Xc.shape)
# fail if win_size exceeds any non-channel dimension
with pytest.raises(ValueError):
structural_similarity(Xc, Yc, win_size=7, multichannel=False)
def test_structural_similarity_nD():
# test 1D through 4D on small random arrays
N = 10
for ndim in range(1, 5):
xsize = [N] * 5
X = (cp.random.rand(*xsize) * 255).astype(cp.uint8)
Y = (cp.random.rand(*xsize) * 255).astype(cp.uint8)
mssim = structural_similarity(X, Y, win_size=3)
assert mssim < 0.05
def test_denoise_tv_chambolle_weighting():
# make sure a specified weight gives consistent results regardless of
# the number of input image dimensions
rstate = cp.random.RandomState(1234)
img2d = astro_gray.copy()
img2d += 0.15 * rstate.standard_normal(img2d.shape)
img2d = cp.clip(img2d, 0, 1)
# generate 4D image by tiling
img4d = cp.tile(img2d[..., None, None], (1, 1, 2, 2))
w = 0.2
denoised_2d = restoration.denoise_tv_chambolle(img2d, weight=w)
denoised_4d = restoration.denoise_tv_chambolle(img4d, weight=w)
assert structural_similarity(denoised_2d, denoised_4d[:, :, 0, 0]) > 0.99
def test_gaussian_mssim_vs_author_ref():
"""
test vs. result from original author's Matlab implementation available at
https://ece.uwaterloo.ca/~z70wang/research/ssim/
Matlab test code:
img1 = imread('camera.png')
img2 = imread('camera_noisy.png')
mssim = structural_similarity_index(img1, img2)
"""
mssim_matlab = 0.327314295673357
mssim = structural_similarity(
cam, cam_noisy, gaussian_weights=True, use_sample_covariance=False
)
assert_almost_equal(mssim, mssim_matlab, decimal=10)
def test_gaussian_structural_similarity_vs_IPOL():
"""Tests vs. imdiff result from the following IPOL article and code:
https://www.ipol.im/pub/art/2011/g_lmii/.
Notes
-----
To generate mssim_IPOL, we need a local copy of cam_noisy:
>>> from skimage import io
>>> io.imsave('/tmp/cam_noisy.png', cam_noisy)
Then, we use the following command:
$ ./imdiff -m mssim <path to camera.png>/camera.png /tmp/cam_noisy.png
Values for current data.camera() calculated by Gregory Lee on Sep, 2020.
Available at:
https://github.com/scikit-image/scikit-image/pull/4913#issuecomment-700653165
"""
mssim_IPOL = 0.357959091663361
mssim = structural_similarity(cam,
cam_noisy,
gaussian_weights=True,
use_sample_covariance=False)
assert_almost_equal(mssim, mssim_IPOL, decimal=3)
def test_gaussian_mssim_and_gradient_vs_Matlab():
# comparison to Matlab implementation of N. Avanaki:
# https://ece.uwaterloo.ca/~nnikvand/Coderep/SHINE%20TOOLBOX/SHINEtoolbox/
# Note: final line of ssim_sens.m was modified to discard image borders
if have_fetch:
ref = np.load(fetch("data/mssim_matlab_output.npz"))
else:
ref = np.load(os.path.join(data_dir, "mssim_matlab_output.npz"))
grad_matlab = ref["grad_matlab"]
mssim_matlab = float(ref["mssim_matlab"])
mssim, grad = structural_similarity(
cam,
cam_noisy,
gaussian_weights=True,
gradient=True,
use_sample_covariance=False,
)
assert_almost_equal(mssim, mssim_matlab, decimal=3)
# check almost equal aside from object borders
assert_array_almost_equal(grad_matlab[5:-5], grad[5:-5])
def test_structural_similarity_image():
N = 100
rstate = cp.random.RandomState(1234)
X = (rstate.rand(N, N) * 255).astype(cp.uint8)
Y = (rstate.rand(N, N) * 255).astype(cp.uint8)
S0 = structural_similarity(X, X, win_size=3)
assert_equal(S0, 1)
S1 = structural_similarity(X, Y, win_size=3)
assert S1 < 0.3
S2 = structural_similarity(X, Y, win_size=11, gaussian_weights=True)
assert S2 < 0.3
mssim0, S3 = structural_similarity(X, Y, full=True)
assert_equal(S3.shape, X.shape)
mssim = structural_similarity(X, Y)
assert_equal(mssim0, mssim)
# structural_similarity of image with itself should be 1.0
assert_equal(structural_similarity(X, X), 1.0)
def test_mssim_vs_legacy():
# check that ssim with default options matches skimage 0.11 result
mssim_skimage_0pt17 = 0.3674518327910367
mssim = structural_similarity(cam, cam_noisy)
assert_almost_equal(mssim, mssim_skimage_0pt17)
def test_mssim_vs_legacy():
# check that ssim with default options matches skimage 0.11 result
mssim_skimage_0pt11 = 0.34192589699605191
mssim = structural_similarity(cam, cam_noisy)
assert_almost_equal(mssim, mssim_skimage_0pt11)
