File size: 14,421 Bytes
36c95ba | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 | from itertools import zip_longest
from typing import Iterator, List, Optional, Tuple, Type, Union
import torch
import torch.nn as nn
import kornia
from kornia.augmentation.base import (
_AugmentationBase,
GeometricAugmentationBase2D,
IntensityAugmentationBase2D,
MixAugmentationBase,
TensorWithTransformMat,
)
from .base import ParamItem, SequentialBase
from .utils import ApplyInverseInterface, InputApplyInverse
__all__ = ["ImageSequential"]
class ImageSequential(SequentialBase):
r"""Sequential for creating kornia image processing pipeline.
Args:
*args : a list of kornia augmentation and image operation modules.
same_on_batch: apply the same transformation across the batch.
If None, it will not overwrite the function-wise settings.
return_transform: if ``True`` return the matrix describing the transformation
applied to each. If None, it will not overwrite the function-wise settings.
keepdim: whether to keep the output shape the same as input (True) or broadcast it
to the batch form (False). If None, it will not overwrite the function-wise settings.
random_apply: randomly select a sublist (order agnostic) of args to
apply transformation.
If int, a fixed number of transformations will be selected.
If (a,), x number of transformations (a <= x <= len(args)) will be selected.
If (a, b), x number of transformations (a <= x <= b) will be selected.
If True, the whole list of args will be processed as a sequence in a random order.
If False, the whole list of args will be processed as a sequence in original order.
.. note::
Transformation matrix returned only considers the transformation applied in ``kornia.augmentation`` module.
Those transformations in ``kornia.geometry`` will not be taken into account.
Examples:
>>> _ = torch.manual_seed(77)
>>> import kornia
>>> input, label = torch.randn(2, 3, 5, 6), torch.tensor([0, 1])
>>> aug_list = ImageSequential(
... kornia.color.BgrToRgb(),
... kornia.augmentation.ColorJitter(0.1, 0.1, 0.1, 0.1, p=1.0),
... kornia.filters.MedianBlur((3, 3)),
... kornia.augmentation.RandomAffine(360, p=1.0),
... kornia.enhance.Invert(),
... kornia.augmentation.RandomMixUp(p=1.0),
... return_transform=True,
... same_on_batch=True,
... random_apply=10,
... )
>>> out, lab = aug_list(input, label=label)
>>> lab
tensor([[0.0000, 1.0000, 0.1214],
[1.0000, 0.0000, 0.1214]])
>>> out[0].shape, out[1].shape
(torch.Size([2, 3, 5, 6]), torch.Size([2, 3, 3]))
Reproduce with provided params.
>>> out2, lab2 = aug_list(input, label=label, params=aug_list._params)
>>> torch.equal(out[0], out2[0]), torch.equal(out[1], out2[1]), torch.equal(lab[1], lab2[1])
(True, True, True)
"""
def __init__(
self,
*args: nn.Module,
same_on_batch: Optional[bool] = None,
return_transform: Optional[bool] = None,
keepdim: Optional[bool] = None,
random_apply: Union[int, bool, Tuple[int, int]] = False,
if_unsupported_ops: str = 'raise'
) -> None:
super().__init__(*args, same_on_batch=same_on_batch, return_transform=return_transform, keepdim=keepdim)
self.random_apply: Union[Tuple[int, int], bool] = self._read_random_apply(random_apply, len(args))
self.return_label: Optional[bool] = None
self.apply_inverse_func: Type[ApplyInverseInterface] = InputApplyInverse
self.if_unsupported_ops = if_unsupported_ops
def _read_random_apply(
self, random_apply: Union[int, bool, Tuple[int, int]], max_length: int
) -> Union[Tuple[int, int], bool]:
"""Process the scenarios for random apply."""
if isinstance(random_apply, (bool,)) and random_apply is False:
random_apply = False
elif isinstance(random_apply, (bool,)) and random_apply is True:
random_apply = (max_length, max_length + 1)
elif isinstance(random_apply, (int,)):
random_apply = (random_apply, random_apply + 1)
elif (
isinstance(random_apply, (tuple,))
and len(random_apply) == 2
and isinstance(random_apply[0], (int,))
and isinstance(random_apply[1], (int,))
):
random_apply = (random_apply[0], random_apply[1] + 1)
elif isinstance(random_apply, (tuple,)) and len(random_apply) == 1 and isinstance(random_apply[0], (int,)):
random_apply = (random_apply[0], max_length + 1)
else:
raise ValueError(f"Non-readable random_apply. Got {random_apply}.")
if random_apply is not False and not (
isinstance(random_apply, (tuple,))
and len(random_apply) == 2
and isinstance(random_apply[0], (int,))
and isinstance(random_apply[0], (int,))
):
raise AssertionError(f"Expect a tuple of (int, int). Got {random_apply}.")
return random_apply
def get_random_forward_sequence(self, with_mix: bool = True) -> Tuple[Iterator[Tuple[str, nn.Module]], bool]:
"""Get a forward sequence when random apply is in need.
Note:
Mix augmentations (e.g. RandomMixUp) will be only applied once even in a random forward.
"""
num_samples = int(torch.randint(*self.random_apply, (1,)).item()) # type: ignore
multinomial_weights = torch.ones((len(self),))
# Mix augmentation can only be applied once per forward
mix_indices = self.get_mix_augmentation_indices(self.named_children())
# kick out the mix augmentations
multinomial_weights[mix_indices] = 0
indices = torch.multinomial(
multinomial_weights,
num_samples,
# enable replacement if non-mix augmentation is less than required
replacement=num_samples > multinomial_weights.sum().item(),
)
mix_added = False
if with_mix and len(mix_indices) != 0:
# Make the selection fair.
if (torch.rand(1) < ((len(mix_indices) + len(indices)) / len(self))).item():
indices[-1] = torch.multinomial((~multinomial_weights.bool()).float(), 1)
indices = indices[torch.randperm(len(indices))]
mix_added = True
return self.get_children_by_indices(indices), mix_added
def get_mix_augmentation_indices(self, named_modules: Iterator[Tuple[str, nn.Module]]) -> List[int]:
"""Get all the mix augmentations since they are label-involved.
Special operations needed for label-involved augmentations.
"""
indices = []
for idx, (_, child) in enumerate(named_modules):
if isinstance(child, (MixAugmentationBase,)):
indices.append(idx)
return indices
def get_forward_sequence(self, params: Optional[List[ParamItem]] = None) -> Iterator[Tuple[str, nn.Module]]:
if params is None:
# Mix augmentation can only be applied once per forward
mix_indices = self.get_mix_augmentation_indices(self.named_children())
if self.random_apply:
return self.get_random_forward_sequence()[0]
if len(mix_indices) > 1:
raise ValueError(
"Multiple mix augmentation is prohibited without enabling random_apply."
f"Detected {len(mix_indices)}."
)
return self.named_children()
return self.get_children_by_params(params)
def apply_to_input(
self,
input: TensorWithTransformMat,
label: Optional[torch.Tensor],
module: Optional[nn.Module],
param: ParamItem,
) -> Tuple[TensorWithTransformMat, Optional[torch.Tensor]]:
if module is None:
module = self.get_submodule(param.name)
return self.apply_inverse_func.apply_trans(input, label, module, param) # type: ignore
def forward_parameters(self, batch_shape: torch.Size) -> List[ParamItem]:
named_modules: Iterator[Tuple[str, nn.Module]] = self.get_forward_sequence()
params: List[ParamItem] = []
mod_param: Union[dict, list]
for name, module in named_modules:
if isinstance(module, (_AugmentationBase, MixAugmentationBase)):
mod_param = module.forward_parameters(batch_shape)
param = ParamItem(name, mod_param)
elif isinstance(module, ImageSequential):
mod_param = module.forward_parameters(batch_shape)
param = ParamItem(name, mod_param)
else:
param = ParamItem(name, None)
params.append(param)
return params
def contains_label_operations(self, params: List[ParamItem]) -> bool:
"""Check if current sequential contains label-involved operations like MixUp."""
for param in params:
if param.name.startswith("RandomMixUp") or param.name.startswith("RandomCutMix"):
return True
return False
def __packup_output__(
self, output: TensorWithTransformMat, label: Optional[torch.Tensor] = None
) -> Union[TensorWithTransformMat, Tuple[TensorWithTransformMat, torch.Tensor]]:
if self.return_label:
return output, label # type: ignore
# Implicitly indicating the label cannot be optional since there is a mix aug
return output
def get_transformation_matrix(
self, input: torch.Tensor, params: Optional[List[ParamItem]] = None,
) -> torch.Tensor:
"""Compute the transformation matrix according to the provided parameters."""
if params is None:
raise NotImplementedError("requires params to be provided.")
named_modules: Iterator[Tuple[str, nn.Module]] = self.get_forward_sequence(params)
res_mat: torch.Tensor = kornia.eye_like(3, input)
for (_, module), param in zip(named_modules, params):
if isinstance(module, (_AugmentationBase, MixAugmentationBase)):
mat = module.compute_transformation(input, param.data) # type: ignore
res_mat = mat @ res_mat
elif isinstance(module, (ImageSequential,)):
mat = module.get_transformation_matrix(input, param.data) # type: ignore
res_mat = mat @ res_mat
return res_mat
def is_intensity_only(self, strict: bool = True) -> bool:
"""Check if all transformations are intensity-based.
Args:
strict: if strict is False, it will allow non-augmentation nn.Modules to be passed.
e.g. `kornia.enhance.AdjustBrightness` will be recognized as non-intensity module
if strict is set to True.
Note: patch processing would break the continuity of labels (e.g. bbounding boxes, masks).
"""
for arg in self.children():
if isinstance(arg, (ImageSequential,)) and not arg.is_intensity_only(strict):
return False
elif isinstance(arg, (ImageSequential,)):
pass
elif isinstance(arg, IntensityAugmentationBase2D):
pass
elif strict:
# disallow non-registered ops if in strict mode
# TODO: add an ops register module
return False
return True
def inverse(
self,
input: torch.Tensor,
params: Optional[List[ParamItem]] = None,
) -> torch.Tensor:
"""Inverse transformation.
Used to inverse a tensor according to the performed transformation by a forward pass, or with respect to
provided parameters.
"""
if params is None:
if self._params is None:
raise ValueError(
"No parameters available for inversing, please run a forward pass first "
"or passing valid params into this function."
)
params = self._params
for (name, module), param in zip_longest(list(self.get_forward_sequence(params))[::-1], params[::-1]):
if isinstance(module, (_AugmentationBase, ImageSequential)):
param = params[name] if name in params else param
else:
param = None
if isinstance(module, IntensityAugmentationBase2D):
pass # Do nothing
elif isinstance(module, ImageSequential) and module.is_intensity_only():
pass # Do nothing
elif isinstance(module, ImageSequential):
input = module.inverse(input, param.data)
elif isinstance(module, (GeometricAugmentationBase2D,)):
input = self.apply_inverse_func.inverse(input, module, param)
else:
pass
# raise NotImplementedError(f"`inverse` is not implemented for {module}.")
return input
def forward( # type: ignore
self,
input: TensorWithTransformMat,
label: Optional[torch.Tensor] = None,
params: Optional[List[ParamItem]] = None,
) -> Union[TensorWithTransformMat, Tuple[TensorWithTransformMat, torch.Tensor]]:
self.clear_state()
if params is None:
if isinstance(input, (tuple, list)):
inp = input[0]
else:
inp = input
_, out_shape = self.autofill_dim(inp, dim_range=(2, 4))
params = self.forward_parameters(out_shape)
if self.return_label is None:
self.return_label = label is not None or self.contains_label_operations(params)
for param in params:
module = self.get_submodule(param.name)
input, label = self.apply_to_input(input, label, module, param=param) # type: ignore
if isinstance(module, (_AugmentationBase, MixAugmentationBase, SequentialBase)):
param = ParamItem(param.name, module._params)
else:
param = ParamItem(param.name, None)
self.update_params(param)
return self.__packup_output__(input, label)
|