def __call__(self, results):
if np.random.rand() > self.prob:
return results
magnitude = random_negative(self.magnitude, self.random_negative_prob)
for key in results.get('img_fields', ['img']):
img = results[key]
img_sharpened = mmcv.adjust_sharpness(img, factor=1 + magnitude)
results[key] = img_sharpened.astype(img.dtype)
return results
def test_adjust_sharpness(self, nb_rand_test=100):
def _adjust_sharpness(img, factor):
# adjust the sharpness of image using
# PIL.ImageEnhance.Sharpness
from PIL.ImageEnhance import Sharpness
from PIL import Image
img = Image.fromarray(img)
sharpened_img = Sharpness(img).enhance(factor)
return np.asarray(sharpened_img)
img = np.array([[0, 128, 255], [1, 127, 254], [2, 129, 253]],
dtype=np.uint8)
img = np.stack([img, img, img], axis=-1)
# test case with invalid type of kernel
with pytest.raises(AssertionError):
mmcv.adjust_sharpness(img, 1., kernel=1.)
# test case with invalid shape of kernel
kernel = np.ones((3, 3, 3))
with pytest.raises(AssertionError):
mmcv.adjust_sharpness(img, 1., kernel=kernel)
# test case with all-zero kernel, factor 0.0
kernel = np.zeros((3, 3))
assert_array_equal(
mmcv.adjust_sharpness(img, 0., kernel=kernel), np.zeros_like(img))
# test case with factor 1.0
assert_array_equal(mmcv.adjust_sharpness(img, 1.), img)
# test adjust_sharpness with randomly sampled images and factors.
for _ in range(nb_rand_test):
img = np.clip(
np.random.uniform(0, 1, (1000, 1200, 3)) * 260, 0,
255).astype(np.uint8)
factor = np.random.uniform()
# Note the gap between PIL.ImageEnhance.Sharpness and
# mmcv.adjust_sharpness mainly comes from the difference ways of
# handling img edges when applying filters
np.testing.assert_allclose(
mmcv.adjust_sharpness(img, factor).astype(np.int32)[1:-1,
1:-1],
_adjust_sharpness(img, factor).astype(np.int32)[1:-1, 1:-1],
rtol=0,
atol=1)
