def adjust_saturation(input: torch.Tensor,
saturation_factor: float) -> torch.Tensor:
r"""Adjust color saturation of an image.
See :class:`~kornia.color.AdjustSaturation` for details.
"""
if not torch.is_tensor(input):
raise TypeError(f"Input type is not a torch.Tensor. Got {type(input)}")
if not isinstance(saturation_factor, float):
raise TypeError(f"The saturation_factor should be a float number.")
# convert the rgb image to hsv
x_hsv: torch.Tensor = rgb_to_hsv(input)
# unpack the hsv values
h, s, v = torch.chunk(x_hsv, chunks=3, dim=-3)
# transform the hue value and appl module
s_out: torch.Tensor = torch.clamp(s * saturation_factor, min=0, max=1)
# pack back back the corrected hue
x_adjusted: torch.Tensor = torch.cat([h, s_out, v], dim=-3)
# convert back to rgb
out: torch.Tensor = hsv_to_rgb(x_adjusted)
return out
def adjust_hue(input: torch.Tensor, hue_factor: float) -> torch.Tensor:
r"""Adjust hue of an image.
See :class:`~kornia.color.AdjustHue` for details.
"""
if not torch.is_tensor(input):
raise TypeError(f"Input type is not a torch.Tensor. Got {type(input)}")
if not isinstance(hue_factor, float) and -0.5 < hue_factor < 0.5:
raise TypeError(
f"The hue_factor should be a float number in the range between"
f" [-0.5, 0.5]. Got {type(hue_factor)}")
# convert the rgb image to hsv
x_hsv: torch.Tensor = rgb_to_hsv(input)
# unpack the hsv values
h, s, v = torch.chunk(x_hsv, chunks=3, dim=-3)
# transform the hue value and appl module
h_out: torch.Tensor = torch.frac(h * hue_factor)
# pack back back the corrected hue
x_adjusted: torch.Tensor = torch.cat([h_out, s, v], dim=-3)
# convert back to rgb
out: torch.Tensor = hsv_to_rgb(x_adjusted)
return out
def transform(self, x):
""" Returns jittered images.
Args:
x (torch.Tensor): observation tensor.
Returns:
torch.Tensor: processed observation tensor.
"""
# check if channel can be devided by three
if x.shape[1] % 3 > 0:
raise ValueError('color jitter is used with stacked RGB images')
# flag for transformation order
is_transforming_rgb_first = np.random.randint(2)
# (batch, C, W, H) -> (batch, stack, 3, W, H)
flat_rgb = x.view(x.shape[0], -1, 3, x.shape[2], x.shape[3])
if is_transforming_rgb_first:
# transform contrast
flat_rgb = self._transform_contrast(flat_rgb)
# (batch, stack, 3, W, H) -> (batch * stack, 3, W, H)
rgb_images = flat_rgb.view(-1, 3, x.shape[2], x.shape[3])
# RGB -> HSV
hsv_images = rgb_to_hsv(rgb_images)
# apply same transformation within the stacked images
# (batch * stack, 3, W, H) -> (batch, stack, 3, W, H)
flat_hsv = hsv_images.view(x.shape[0], -1, 3, x.shape[2], x.shape[3])
# transform hue
flat_hsv = self._transform_hue(flat_hsv)
# transform saturate
flat_hsv = self._transform_saturate(flat_hsv)
# transform brightness
flat_hsv = self._transform_brightness(flat_hsv)
# (batch, stack, 3, W, H) -> (batch * stack, 3, W, H)
hsv_images = flat_hsv.view(-1, 3, x.shape[2], x.shape[3])
# HSV -> RGB
rgb_images = hsv_to_rgb(hsv_images)
# (batch * stack, 3, W, H) -> (batch, stack, 3, W, H)
flat_rgb = rgb_images.view(x.shape[0], -1, 3, x.shape[2], x.shape[3])
if not is_transforming_rgb_first:
# transform contrast
flat_rgb = self._transform_contrast(flat_rgb)
return flat_rgb.view(*x.shape)
def adjust_saturation(
input: torch.Tensor,
saturation_factor: Union[float, torch.Tensor]) -> torch.Tensor:
r"""Adjust color saturation of an image.
The input image is expected to be an RGB image in the range of [0, 1].
Args:
input (torch.Tensor): Image/Tensor to be adjusted in the shape of :math:`(*, 3, H, W)`.
saturation_factor (Union[float, torch.Tensor]): How much to adjust the saturation. 0 will give a black
and white image, 1 will give the original image while 2 will enhance the saturation
by a factor of 2.
Return:
torch.Tensor: Adjusted image in the shape of :math:`(*, 3, H, W)`.
Example:
>>> x = torch.ones(1, 3, 3, 3)
>>> adjust_saturation(x, 2.)
tensor([[[[1., 1., 1.],
[1., 1., 1.],
[1., 1., 1.]],
<BLANKLINE>
[[1., 1., 1.],
[1., 1., 1.],
[1., 1., 1.]],
<BLANKLINE>
[[1., 1., 1.],
[1., 1., 1.],
[1., 1., 1.]]]])
>>> x = torch.ones(2, 3, 3, 3)
>>> y = torch.ones(2)
>>> out = adjust_saturation(x, y)
>>> torch.nn.functional.mse_loss(x, out)
tensor(0.)
"""
# convert the rgb image to hsv
x_hsv: torch.Tensor = rgb_to_hsv(input)
# perform the conversion
x_adjusted: torch.Tensor = adjust_saturation_raw(x_hsv, saturation_factor)
# convert back to rgb
out: torch.Tensor = hsv_to_rgb(x_adjusted)
return out
def adjust_saturation(input: torch.Tensor, saturation_factor: Union[float, torch.Tensor]) -> torch.Tensor:
r"""Adjust color saturation of an image.
See :class:`~kornia.color.AdjustSaturation` for details.
"""
# convert the rgb image to hsv
x_hsv: torch.Tensor = rgb_to_hsv(input)
# perform the conversion
x_adjusted: torch.Tensor = adjust_saturation_raw(x_hsv, saturation_factor)
# convert back to rgb
out: torch.Tensor = hsv_to_rgb(x_adjusted)
return out
def adjust_hue(input: torch.Tensor,
hue_factor: Union[float, torch.Tensor]) -> torch.Tensor:
r"""Adjust hue of an image.
The input image is expected to be an RGB image in the range of [0, 1].
Args:
input (torch.Tensor): Image to be adjusted in the shape of :math:`(*, 3, H, W)`.
hue_factor (Union[float, torch.Tensor]): How much to shift the hue channel. Should be in [-PI, PI]. PI
and -PI give complete reversal of hue channel in HSV space in positive and negative
direction respectively. 0 means no shift. Therefore, both -PI and PI will give an
image with complementary colors while 0 gives the original image.
Return:
torch.Tensor: Adjusted image in the shape of :math:`(*, 3, H, W)`.
Example:
>>> x = torch.ones(1, 3, 3, 3)
>>> adjust_hue(x, 3.141516)
tensor([[[[1., 1., 1.],
[1., 1., 1.],
[1., 1., 1.]],
<BLANKLINE>
[[1., 1., 1.],
[1., 1., 1.],
[1., 1., 1.]],
<BLANKLINE>
[[1., 1., 1.],
[1., 1., 1.],
[1., 1., 1.]]]])
>>> x = torch.ones(2, 3, 3, 3)
>>> y = torch.ones(2) * 3.141516
>>> adjust_hue(x, y).shape
torch.Size([2, 3, 3, 3])
"""
# convert the rgb image to hsv
x_hsv: torch.Tensor = rgb_to_hsv(input)
# perform the conversion
x_adjusted: torch.Tensor = adjust_hue_raw(x_hsv, hue_factor)
# convert back to rgb
out: torch.Tensor = hsv_to_rgb(x_adjusted)
return out
def adjust_saturation(
input: torch.Tensor,
saturation_factor: Union[float, torch.Tensor]) -> torch.Tensor:
r"""Adjust color saturation of an image.
See :class:`~kornia.color.AdjustSaturation` for details.
"""
if not torch.is_tensor(input):
raise TypeError(f"Input type is not a torch.Tensor. Got {type(input)}")
if not isinstance(saturation_factor, (
float,
torch.Tensor,
)):
raise TypeError(
f"The saturation_factor should be a float number or torch.Tensor."
f"Got {type(saturation_factor)}")
if isinstance(saturation_factor, float):
saturation_factor = torch.tensor([saturation_factor])
saturation_factor = saturation_factor.to(input.device).to(input.dtype)
if (saturation_factor < 0).any():
raise ValueError(
f"Saturation factor must be non-negative. Got {saturation_factor}")
for _ in input.shape[1:]:
saturation_factor = torch.unsqueeze(saturation_factor, dim=-1)
# convert the rgb image to hsv
x_hsv: torch.Tensor = rgb_to_hsv(input)
# unpack the hsv values
h, s, v = torch.chunk(x_hsv, chunks=3, dim=-3)
# transform the hue value and appl module
s_out: torch.Tensor = torch.clamp(s * saturation_factor, min=0, max=1)
# pack back back the corrected hue
x_adjusted: torch.Tensor = torch.cat([h, s_out, v], dim=-3)
# convert back to rgb
out: torch.Tensor = hsv_to_rgb(x_adjusted)
return out
def adjust_hue(input: torch.Tensor,
hue_factor: Union[float, torch.Tensor]) -> torch.Tensor:
r"""Adjust hue of an image.
See :class:`~kornia.color.AdjustHue` for details.
"""
if not torch.is_tensor(input):
raise TypeError(f"Input type is not a torch.Tensor. Got {type(input)}")
if not isinstance(hue_factor, (float, torch.Tensor)):
raise TypeError(
f"The hue_factor should be a float number or torch.Tensor in the range between"
f" [-PI, PI]. Got {type(hue_factor)}")
if isinstance(hue_factor, float):
hue_factor = torch.tensor([hue_factor])
hue_factor = hue_factor.to(input.device).to(input.dtype)
if ((hue_factor < -pi) | (hue_factor > pi)).any():
raise ValueError(
f"Hue-factor must be in the range [-PI, PI]. Got {hue_factor}")
for _ in input.shape[1:]:
hue_factor = torch.unsqueeze(hue_factor, dim=-1)
# convert the rgb image to hsv
x_hsv: torch.Tensor = rgb_to_hsv(input)
# unpack the hsv values
h, s, v = torch.chunk(x_hsv, chunks=3, dim=-3)
# transform the hue value and appl module
divisor: float = 2 * pi.item()
h_out: torch.Tensor = torch.fmod(h + hue_factor, divisor)
# pack back back the corrected hue
x_adjusted: torch.Tensor = torch.cat([h_out, s, v], dim=-3)
# convert back to rgb
out: torch.Tensor = hsv_to_rgb(x_adjusted)
return out
def adjust_hue(input: torch.Tensor,
hue_factor: Union[float, torch.Tensor]) -> torch.Tensor:
r"""Adjust hue of an image.
.. image:: _static/img/adjust_hue.png
The input image is expected to be an RGB image in the range of [0, 1].
Args:
input: Image to be adjusted in the shape of :math:`(*, 3, H, W)`.
hue_factor: How much to shift the hue channel. Should be in [-PI, PI]. PI
and -PI give complete reversal of hue channel in HSV space in positive and negative
direction respectively. 0 means no shift. Therefore, both -PI and PI will give an
image with complementary colors while 0 gives the original image.
Return:
Adjusted image in the shape of :math:`(*, 3, H, W)`.
.. note::
See a working example `here <https://kornia-tutorials.readthedocs.io/en/latest/
image_enhancement.html>`__.
Example:
>>> x = torch.ones(1, 3, 2, 2)
>>> adjust_hue(x, 3.141516).shape
torch.Size([1, 3, 2, 2])
>>> x = torch.ones(2, 3, 3, 3)
>>> y = torch.ones(2) * 3.141516
>>> adjust_hue(x, y).shape
torch.Size([2, 3, 3, 3])
"""
# convert the rgb image to hsv
x_hsv: torch.Tensor = rgb_to_hsv(input)
# perform the conversion
x_adjusted: torch.Tensor = adjust_hue_raw(x_hsv, hue_factor)
# convert back to rgb
out: torch.Tensor = hsv_to_rgb(x_adjusted)
return out
def adjust_saturation(
input: torch.Tensor,
saturation_factor: Union[float, torch.Tensor]) -> torch.Tensor:
r"""Adjust color saturation of an image.
.. image:: _static/img/adjust_saturation.png
The input image is expected to be an RGB image in the range of [0, 1].
Args:
input: Image/Tensor to be adjusted in the shape of :math:`(*, 3, H, W)`.
saturation_factor: How much to adjust the saturation. 0 will give a black
and white image, 1 will give the original image while 2 will enhance the saturation by a factor of 2.
Return:
Adjusted image in the shape of :math:`(*, 3, H, W)`.
.. note::
See a working example `here <https://kornia-tutorials.readthedocs.io/en/latest/
image_enhancement.html>`__.
Example:
>>> x = torch.ones(1, 3, 3, 3)
>>> adjust_saturation(x, 2.).shape
torch.Size([1, 3, 3, 3])
>>> x = torch.ones(2, 3, 3, 3)
>>> y = torch.tensor([1., 2.])
>>> adjust_saturation(x, y).shape
torch.Size([2, 3, 3, 3])
"""
# convert the rgb image to hsv
x_hsv: torch.Tensor = rgb_to_hsv(input)
# perform the conversion
x_adjusted: torch.Tensor = adjust_saturation_raw(x_hsv, saturation_factor)
# convert back to rgb
out: torch.Tensor = hsv_to_rgb(x_adjusted)
return out
