functional.py 48.3 KB
Newer Older
1
import math
2
3
import numbers
import warnings
4
from enum import Enum
vfdev's avatar
vfdev committed
5
from typing import Any, Optional
6
7

import numpy as np
vfdev's avatar
vfdev committed
8
from PIL import Image
9
10
11

import torch
from torch import Tensor
vfdev's avatar
vfdev committed
12
from torch.jit.annotations import List, Tuple
13

14
15
16
17
18
try:
    import accimage
except ImportError:
    accimage = None

19
20
21
from . import functional_pil as F_pil
from . import functional_tensor as F_t

22

23
class InterpolationMode(Enum):
24
25
26
27
28
29
30
31
32
33
34
35
    """Interpolation modes
    """
    NEAREST = "nearest"
    BILINEAR = "bilinear"
    BICUBIC = "bicubic"
    # For PIL compatibility
    BOX = "box"
    HAMMING = "hamming"
    LANCZOS = "lanczos"


# TODO: Once torchscript supports Enums with staticmethod
36
37
# this can be put into InterpolationMode as staticmethod
def _interpolation_modes_from_int(i: int) -> InterpolationMode:
38
    inverse_modes_mapping = {
39
40
41
42
43
44
        0: InterpolationMode.NEAREST,
        2: InterpolationMode.BILINEAR,
        3: InterpolationMode.BICUBIC,
        4: InterpolationMode.BOX,
        5: InterpolationMode.HAMMING,
        1: InterpolationMode.LANCZOS,
45
46
47
48
49
    }
    return inverse_modes_mapping[i]


pil_modes_mapping = {
50
51
52
53
54
55
    InterpolationMode.NEAREST: 0,
    InterpolationMode.BILINEAR: 2,
    InterpolationMode.BICUBIC: 3,
    InterpolationMode.BOX: 4,
    InterpolationMode.HAMMING: 5,
    InterpolationMode.LANCZOS: 1,
56
57
}

vfdev's avatar
vfdev committed
58
_is_pil_image = F_pil._is_pil_image
vfdev's avatar
vfdev committed
59
_parse_fill = F_pil._parse_fill
vfdev's avatar
vfdev committed
60
61
62
63
64
65
66


def _get_image_size(img: Tensor) -> List[int]:
    """Returns image sizea as (w, h)
    """
    if isinstance(img, torch.Tensor):
        return F_t._get_image_size(img)
67

vfdev's avatar
vfdev committed
68
    return F_pil._get_image_size(img)
69

vfdev's avatar
vfdev committed
70

71
72
73
74
75
76
77
def _get_image_num_channels(img: Tensor) -> int:
    if isinstance(img, torch.Tensor):
        return F_t._get_image_num_channels(img)

    return F_pil._get_image_num_channels(img)


vfdev's avatar
vfdev committed
78
79
@torch.jit.unused
def _is_numpy(img: Any) -> bool:
80
81
82
    return isinstance(img, np.ndarray)


vfdev's avatar
vfdev committed
83
84
@torch.jit.unused
def _is_numpy_image(img: Any) -> bool:
85
    return img.ndim in {2, 3}
86
87
88
89
90
91
92
93
94
95
96
97
98


def to_tensor(pic):
    """Convert a ``PIL Image`` or ``numpy.ndarray`` to tensor.

    See ``ToTensor`` for more details.

    Args:
        pic (PIL Image or numpy.ndarray): Image to be converted to tensor.

    Returns:
        Tensor: Converted image.
    """
vfdev's avatar
vfdev committed
99
    if not(F_pil._is_pil_image(pic) or _is_numpy(pic)):
100
101
        raise TypeError('pic should be PIL Image or ndarray. Got {}'.format(type(pic)))

102
103
104
    if _is_numpy(pic) and not _is_numpy_image(pic):
        raise ValueError('pic should be 2/3 dimensional. Got {} dimensions.'.format(pic.ndim))

105
106
    if isinstance(pic, np.ndarray):
        # handle numpy array
surgan12's avatar
surgan12 committed
107
108
109
        if pic.ndim == 2:
            pic = pic[:, :, None]

110
        img = torch.from_numpy(pic.transpose((2, 0, 1))).contiguous()
111
        # backward compatibility
112
113
114
115
        if isinstance(img, torch.ByteTensor):
            return img.float().div(255)
        else:
            return img
116
117
118
119
120
121
122
123
124
125
126

    if accimage is not None and isinstance(pic, accimage.Image):
        nppic = np.zeros([pic.channels, pic.height, pic.width], dtype=np.float32)
        pic.copyto(nppic)
        return torch.from_numpy(nppic)

    # handle PIL Image
    if pic.mode == 'I':
        img = torch.from_numpy(np.array(pic, np.int32, copy=False))
    elif pic.mode == 'I;16':
        img = torch.from_numpy(np.array(pic, np.int16, copy=False))
127
128
    elif pic.mode == 'F':
        img = torch.from_numpy(np.array(pic, np.float32, copy=False))
129
130
    elif pic.mode == '1':
        img = 255 * torch.from_numpy(np.array(pic, np.uint8, copy=False))
131
132
    else:
        img = torch.ByteTensor(torch.ByteStorage.from_buffer(pic.tobytes()))
133
134

    img = img.view(pic.size[1], pic.size[0], len(pic.getbands()))
135
    # put it from HWC to CHW format
136
    img = img.permute((2, 0, 1)).contiguous()
137
138
139
140
141
142
    if isinstance(img, torch.ByteTensor):
        return img.float().div(255)
    else:
        return img


143
144
145
def pil_to_tensor(pic):
    """Convert a ``PIL Image`` to a tensor of the same type.

vfdev's avatar
vfdev committed
146
    See :class:`~torchvision.transforms.PILToTensor` for more details.
147
148
149
150
151
152
153

    Args:
        pic (PIL Image): Image to be converted to tensor.

    Returns:
        Tensor: Converted image.
    """
154
    if not F_pil._is_pil_image(pic):
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
        raise TypeError('pic should be PIL Image. Got {}'.format(type(pic)))

    if accimage is not None and isinstance(pic, accimage.Image):
        nppic = np.zeros([pic.channels, pic.height, pic.width], dtype=np.float32)
        pic.copyto(nppic)
        return torch.as_tensor(nppic)

    # handle PIL Image
    img = torch.as_tensor(np.asarray(pic))
    img = img.view(pic.size[1], pic.size[0], len(pic.getbands()))
    # put it from HWC to CHW format
    img = img.permute((2, 0, 1))
    return img


170
171
172
173
174
175
176
177
def convert_image_dtype(image: torch.Tensor, dtype: torch.dtype = torch.float) -> torch.Tensor:
    """Convert a tensor image to the given ``dtype`` and scale the values accordingly

    Args:
        image (torch.Tensor): Image to be converted
        dtype (torch.dtype): Desired data type of the output

    Returns:
vfdev's avatar
vfdev committed
178
        Tensor: Converted image
179
180
181
182
183
184
185
186
187
188
189
190

    .. note::

        When converting from a smaller to a larger integer ``dtype`` the maximum values are **not** mapped exactly.
        If converted back and forth, this mismatch has no effect.

    Raises:
        RuntimeError: When trying to cast :class:`torch.float32` to :class:`torch.int32` or :class:`torch.int64` as
            well as for trying to cast :class:`torch.float64` to :class:`torch.int64`. These conversions might lead to
            overflow errors since the floating point ``dtype`` cannot store consecutive integers over the whole range
            of the integer ``dtype``.
    """
191
192
193
194
    if not isinstance(image, torch.Tensor):
        raise TypeError('Input img should be Tensor Image')

    return F_t.convert_image_dtype(image, dtype)
195
196


197
198
199
def to_pil_image(pic, mode=None):
    """Convert a tensor or an ndarray to PIL Image.

200
    See :class:`~torchvision.transforms.ToPILImage` for more details.
201
202
203
204
205

    Args:
        pic (Tensor or numpy.ndarray): Image to be converted to PIL Image.
        mode (`PIL.Image mode`_): color space and pixel depth of input data (optional).

206
    .. _PIL.Image mode: https://pillow.readthedocs.io/en/latest/handbook/concepts.html#concept-modes
207
208
209
210

    Returns:
        PIL Image: Image converted to PIL Image.
    """
Varun Agrawal's avatar
Varun Agrawal committed
211
    if not(isinstance(pic, torch.Tensor) or isinstance(pic, np.ndarray)):
212
213
        raise TypeError('pic should be Tensor or ndarray. Got {}.'.format(type(pic)))

Varun Agrawal's avatar
Varun Agrawal committed
214
215
216
217
218
219
    elif isinstance(pic, torch.Tensor):
        if pic.ndimension() not in {2, 3}:
            raise ValueError('pic should be 2/3 dimensional. Got {} dimensions.'.format(pic.ndimension()))

        elif pic.ndimension() == 2:
            # if 2D image, add channel dimension (CHW)
Surgan Jandial's avatar
Surgan Jandial committed
220
            pic = pic.unsqueeze(0)
Varun Agrawal's avatar
Varun Agrawal committed
221

222
223
224
225
        # check number of channels
        if pic.shape[-3] > 4:
            raise ValueError('pic should not have > 4 channels. Got {} channels.'.format(pic.shape[-3]))

Varun Agrawal's avatar
Varun Agrawal committed
226
227
228
229
230
231
232
233
    elif isinstance(pic, np.ndarray):
        if pic.ndim not in {2, 3}:
            raise ValueError('pic should be 2/3 dimensional. Got {} dimensions.'.format(pic.ndim))

        elif pic.ndim == 2:
            # if 2D image, add channel dimension (HWC)
            pic = np.expand_dims(pic, 2)

234
235
236
237
        # check number of channels
        if pic.shape[-1] > 4:
            raise ValueError('pic should not have > 4 channels. Got {} channels.'.format(pic.shape[-1]))

238
    npimg = pic
Varun Agrawal's avatar
Varun Agrawal committed
239
    if isinstance(pic, torch.Tensor):
240
241
242
        if pic.is_floating_point() and mode != 'F':
            pic = pic.mul(255).byte()
        npimg = np.transpose(pic.cpu().numpy(), (1, 2, 0))
243
244
245
246
247
248
249
250
251
252

    if not isinstance(npimg, np.ndarray):
        raise TypeError('Input pic must be a torch.Tensor or NumPy ndarray, ' +
                        'not {}'.format(type(npimg)))

    if npimg.shape[2] == 1:
        expected_mode = None
        npimg = npimg[:, :, 0]
        if npimg.dtype == np.uint8:
            expected_mode = 'L'
vfdev's avatar
vfdev committed
253
        elif npimg.dtype == np.int16:
254
            expected_mode = 'I;16'
vfdev's avatar
vfdev committed
255
        elif npimg.dtype == np.int32:
256
257
258
259
260
261
262
263
            expected_mode = 'I'
        elif npimg.dtype == np.float32:
            expected_mode = 'F'
        if mode is not None and mode != expected_mode:
            raise ValueError("Incorrect mode ({}) supplied for input type {}. Should be {}"
                             .format(mode, np.dtype, expected_mode))
        mode = expected_mode

surgan12's avatar
surgan12 committed
264
265
266
267
268
269
270
271
    elif npimg.shape[2] == 2:
        permitted_2_channel_modes = ['LA']
        if mode is not None and mode not in permitted_2_channel_modes:
            raise ValueError("Only modes {} are supported for 2D inputs".format(permitted_2_channel_modes))

        if mode is None and npimg.dtype == np.uint8:
            mode = 'LA'

272
    elif npimg.shape[2] == 4:
surgan12's avatar
surgan12 committed
273
        permitted_4_channel_modes = ['RGBA', 'CMYK', 'RGBX']
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
        if mode is not None and mode not in permitted_4_channel_modes:
            raise ValueError("Only modes {} are supported for 4D inputs".format(permitted_4_channel_modes))

        if mode is None and npimg.dtype == np.uint8:
            mode = 'RGBA'
    else:
        permitted_3_channel_modes = ['RGB', 'YCbCr', 'HSV']
        if mode is not None and mode not in permitted_3_channel_modes:
            raise ValueError("Only modes {} are supported for 3D inputs".format(permitted_3_channel_modes))
        if mode is None and npimg.dtype == np.uint8:
            mode = 'RGB'

    if mode is None:
        raise TypeError('Input type {} is not supported'.format(npimg.dtype))

    return Image.fromarray(npimg, mode=mode)


292
def normalize(tensor: Tensor, mean: List[float], std: List[float], inplace: bool = False) -> Tensor:
293
294
    """Normalize a tensor image with mean and standard deviation.

295
    .. note::
surgan12's avatar
surgan12 committed
296
        This transform acts out of place by default, i.e., it does not mutates the input tensor.
297

298
    See :class:`~torchvision.transforms.Normalize` for more details.
299
300

    Args:
301
        tensor (Tensor): Tensor image of size (C, H, W) or (B, C, H, W) to be normalized.
302
        mean (sequence): Sequence of means for each channel.
303
        std (sequence): Sequence of standard deviations for each channel.
304
        inplace(bool,optional): Bool to make this operation inplace.
305
306
307
308

    Returns:
        Tensor: Normalized Tensor image.
    """
309
310
    if not isinstance(tensor, torch.Tensor):
        raise TypeError('Input tensor should be a torch tensor. Got {}.'.format(type(tensor)))
311

312
313
    if tensor.ndim < 3:
        raise ValueError('Expected tensor to be a tensor image of size (..., C, H, W). Got tensor.size() = '
314
                         '{}.'.format(tensor.size()))
315

surgan12's avatar
surgan12 committed
316
317
318
    if not inplace:
        tensor = tensor.clone()

319
320
321
    dtype = tensor.dtype
    mean = torch.as_tensor(mean, dtype=dtype, device=tensor.device)
    std = torch.as_tensor(std, dtype=dtype, device=tensor.device)
322
323
    if (std == 0).any():
        raise ValueError('std evaluated to zero after conversion to {}, leading to division by zero.'.format(dtype))
324
    if mean.ndim == 1:
325
        mean = mean.view(-1, 1, 1)
326
    if std.ndim == 1:
327
        std = std.view(-1, 1, 1)
328
    tensor.sub_(mean).div_(std)
329
    return tensor
330
331


332
def resize(img: Tensor, size: List[int], interpolation: InterpolationMode = InterpolationMode.BILINEAR) -> Tensor:
vfdev's avatar
vfdev committed
333
334
335
    r"""Resize the input image to the given size.
    The image can be a PIL Image or a torch Tensor, in which case it is expected
    to have [..., H, W] shape, where ... means an arbitrary number of leading dimensions
336
337

    Args:
vfdev's avatar
vfdev committed
338
        img (PIL Image or Tensor): Image to be resized.
339
340
        size (sequence or int): Desired output size. If size is a sequence like
            (h, w), the output size will be matched to this. If size is an int,
Vitaliy Chiley's avatar
Vitaliy Chiley committed
341
            the smaller edge of the image will be matched to this number maintaining
342
            the aspect ratio. i.e, if height > width, then image will be rescaled to
vfdev's avatar
vfdev committed
343
            :math:`\left(\text{size} \times \frac{\text{height}}{\text{width}}, \text{size}\right)`.
344
            In torchscript mode size as single int is not supported, use a tuple or
vfdev's avatar
vfdev committed
345
            list of length 1: ``[size, ]``.
346
347
348
349
        interpolation (InterpolationMode): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationMode`.
            Default is ``InterpolationMode.BILINEAR``. If input is Tensor, only ``InterpolationMode.NEAREST``,
            ``InterpolationMode.BILINEAR`` and ``InterpolationMode.BICUBIC`` are supported.
350
            For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
351
352

    Returns:
vfdev's avatar
vfdev committed
353
        PIL Image or Tensor: Resized image.
354
    """
355
356
357
    # Backward compatibility with integer value
    if isinstance(interpolation, int):
        warnings.warn(
358
359
            "Argument interpolation should be of type InterpolationMode instead of int. "
            "Please, use InterpolationMode enum."
360
361
362
        )
        interpolation = _interpolation_modes_from_int(interpolation)

363
364
    if not isinstance(interpolation, InterpolationMode):
        raise TypeError("Argument interpolation should be a InterpolationMode")
365

vfdev's avatar
vfdev committed
366
    if not isinstance(img, torch.Tensor):
367
368
        pil_interpolation = pil_modes_mapping[interpolation]
        return F_pil.resize(img, size=size, interpolation=pil_interpolation)
vfdev's avatar
vfdev committed
369

370
    return F_t.resize(img, size=size, interpolation=interpolation.value)
371
372
373
374
375
376
377
378


def scale(*args, **kwargs):
    warnings.warn("The use of the transforms.Scale transform is deprecated, " +
                  "please use transforms.Resize instead.")
    return resize(*args, **kwargs)


379
380
381
382
def pad(img: Tensor, padding: List[int], fill: int = 0, padding_mode: str = "constant") -> Tensor:
    r"""Pad the given image on all sides with the given "pad" value.
    The image can be a PIL Image or a torch Tensor, in which case it is expected
    to have [..., H, W] shape, where ... means an arbitrary number of leading dimensions
383
384

    Args:
385
386
        img (PIL Image or Tensor): Image to be padded.
        padding (int or tuple or list): Padding on each border. If a single int is provided this
387
388
            is used to pad all borders. If tuple of length 2 is provided this is the padding
            on left/right and top/bottom respectively. If a tuple of length 4 is provided
389
390
391
392
            this is the padding for the left, top, right and bottom borders respectively.
            In torchscript mode padding as single int is not supported, use a tuple or
            list of length 1: ``[padding, ]``.
        fill (int or str or tuple): Pixel fill value for constant fill. Default is 0. If a tuple of
393
            length 3, it is used to fill R, G, B channels respectively.
394
            This value is only used when the padding_mode is constant. Only int value is supported for Tensors.
395
        padding_mode: Type of padding. Should be: constant, edge, reflect or symmetric. Default is constant.
vfdev's avatar
vfdev committed
396
            Mode symmetric is not yet supported for Tensor inputs.
397
398
399
400
401
402
403
404
405
406
407
408
409
410

            - constant: pads with a constant value, this value is specified with fill

            - edge: pads with the last value on the edge of the image

            - reflect: pads with reflection of image (without repeating the last value on the edge)

                       padding [1, 2, 3, 4] with 2 elements on both sides in reflect mode
                       will result in [3, 2, 1, 2, 3, 4, 3, 2]

            - symmetric: pads with reflection of image (repeating the last value on the edge)

                         padding [1, 2, 3, 4] with 2 elements on both sides in symmetric mode
                         will result in [2, 1, 1, 2, 3, 4, 4, 3]
411
412

    Returns:
413
        PIL Image or Tensor: Padded image.
414
    """
415
416
    if not isinstance(img, torch.Tensor):
        return F_pil.pad(img, padding=padding, fill=fill, padding_mode=padding_mode)
417

418
    return F_t.pad(img, padding=padding, fill=fill, padding_mode=padding_mode)
419
420


vfdev's avatar
vfdev committed
421
422
423
424
425
def crop(img: Tensor, top: int, left: int, height: int, width: int) -> Tensor:
    """Crop the given image at specified location and output size.
    The image can be a PIL Image or a Tensor, in which case it is expected
    to have [..., H, W] shape, where ... means an arbitrary number of leading
    dimensions
426

427
    Args:
vfdev's avatar
vfdev committed
428
        img (PIL Image or Tensor): Image to be cropped. (0,0) denotes the top left corner of the image.
429
430
431
432
        top (int): Vertical component of the top left corner of the crop box.
        left (int): Horizontal component of the top left corner of the crop box.
        height (int): Height of the crop box.
        width (int): Width of the crop box.
433

434
    Returns:
vfdev's avatar
vfdev committed
435
        PIL Image or Tensor: Cropped image.
436
437
    """

vfdev's avatar
vfdev committed
438
439
    if not isinstance(img, torch.Tensor):
        return F_pil.crop(img, top, left, height, width)
440

vfdev's avatar
vfdev committed
441
    return F_t.crop(img, top, left, height, width)
442

vfdev's avatar
vfdev committed
443
444
445
446
447

def center_crop(img: Tensor, output_size: List[int]) -> Tensor:
    """Crops the given image at the center.
    The image can be a PIL Image or a Tensor, in which case it is expected
    to have [..., H, W] shape, where ... means an arbitrary number of leading dimensions
448

449
    Args:
vfdev's avatar
vfdev committed
450
451
452
453
        img (PIL Image or Tensor): Image to be cropped.
        output_size (sequence or int): (height, width) of the crop box. If int or sequence with single int
            it is used for both directions.

454
    Returns:
vfdev's avatar
vfdev committed
455
        PIL Image or Tensor: Cropped image.
456
    """
457
458
    if isinstance(output_size, numbers.Number):
        output_size = (int(output_size), int(output_size))
vfdev's avatar
vfdev committed
459
460
461
462
    elif isinstance(output_size, (tuple, list)) and len(output_size) == 1:
        output_size = (output_size[0], output_size[0])

    image_width, image_height = _get_image_size(img)
463
    crop_height, crop_width = output_size
vfdev's avatar
vfdev committed
464
465
466
467
468
469
470
471
472

    # crop_top = int(round((image_height - crop_height) / 2.))
    # Result can be different between python func and scripted func
    # Temporary workaround:
    crop_top = int((image_height - crop_height + 1) * 0.5)
    # crop_left = int(round((image_width - crop_width) / 2.))
    # Result can be different between python func and scripted func
    # Temporary workaround:
    crop_left = int((image_width - crop_width + 1) * 0.5)
473
    return crop(img, crop_top, crop_left, crop_height, crop_width)
474
475


476
def resized_crop(
477
        img: Tensor, top: int, left: int, height: int, width: int, size: List[int],
478
        interpolation: InterpolationMode = InterpolationMode.BILINEAR
479
480
481
482
) -> Tensor:
    """Crop the given image and resize it to desired size.
    The image can be a PIL Image or a Tensor, in which case it is expected
    to have [..., H, W] shape, where ... means an arbitrary number of leading dimensions
483

484
    Notably used in :class:`~torchvision.transforms.RandomResizedCrop`.
485
486

    Args:
487
        img (PIL Image or Tensor): Image to be cropped. (0,0) denotes the top left corner of the image.
488
489
490
491
        top (int): Vertical component of the top left corner of the crop box.
        left (int): Horizontal component of the top left corner of the crop box.
        height (int): Height of the crop box.
        width (int): Width of the crop box.
492
        size (sequence or int): Desired output size. Same semantics as ``resize``.
493
494
495
496
        interpolation (InterpolationMode): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationMode`.
            Default is ``InterpolationMode.BILINEAR``. If input is Tensor, only ``InterpolationMode.NEAREST``,
            ``InterpolationMode.BILINEAR`` and ``InterpolationMode.BICUBIC`` are supported.
497
498
            For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.

499
    Returns:
500
        PIL Image or Tensor: Cropped image.
501
    """
502
    img = crop(img, top, left, height, width)
503
504
505
506
    img = resize(img, size, interpolation)
    return img


507
def hflip(img: Tensor) -> Tensor:
vfdev's avatar
vfdev committed
508
    """Horizontally flip the given PIL Image or Tensor.
509
510

    Args:
vfdev's avatar
vfdev committed
511
        img (PIL Image or Tensor): Image to be flipped. If img
512
513
514
            is a Tensor, it is expected to be in [..., H, W] format,
            where ... means it can have an arbitrary number of trailing
            dimensions.
515
516

    Returns:
vfdev's avatar
vfdev committed
517
        PIL Image or Tensor:  Horizontally flipped image.
518
    """
519
520
    if not isinstance(img, torch.Tensor):
        return F_pil.hflip(img)
521

522
    return F_t.hflip(img)
523
524


525
526
527
def _get_perspective_coeffs(
        startpoints: List[List[int]], endpoints: List[List[int]]
) -> List[float]:
528
529
    """Helper function to get the coefficients (a, b, c, d, e, f, g, h) for the perspective transforms.

Vitaliy Chiley's avatar
Vitaliy Chiley committed
530
    In Perspective Transform each pixel (x, y) in the original image gets transformed as,
531
532
533
     (x, y) -> ( (ax + by + c) / (gx + hy + 1), (dx + ey + f) / (gx + hy + 1) )

    Args:
534
535
536
537
538
        startpoints (list of list of ints): List containing four lists of two integers corresponding to four corners
            ``[top-left, top-right, bottom-right, bottom-left]`` of the original image.
        endpoints (list of list of ints): List containing four lists of two integers corresponding to four corners
            ``[top-left, top-right, bottom-right, bottom-left]`` of the transformed image.

539
540
541
    Returns:
        octuple (a, b, c, d, e, f, g, h) for transforming each pixel.
    """
542
543
544
545
546
    a_matrix = torch.zeros(2 * len(startpoints), 8, dtype=torch.float)

    for i, (p1, p2) in enumerate(zip(endpoints, startpoints)):
        a_matrix[2 * i, :] = torch.tensor([p1[0], p1[1], 1, 0, 0, 0, -p2[0] * p1[0], -p2[0] * p1[1]])
        a_matrix[2 * i + 1, :] = torch.tensor([0, 0, 0, p1[0], p1[1], 1, -p2[1] * p1[0], -p2[1] * p1[1]])
547

548
549
    b_matrix = torch.tensor(startpoints, dtype=torch.float).view(8)
    res = torch.lstsq(b_matrix, a_matrix)[0]
550

551
552
    output: List[float] = res.squeeze(1).tolist()
    return output
553
554


555
556
557
558
def perspective(
        img: Tensor,
        startpoints: List[List[int]],
        endpoints: List[List[int]],
559
        interpolation: InterpolationMode = InterpolationMode.BILINEAR,
560
        fill: Optional[List[float]] = None
561
562
563
564
) -> Tensor:
    """Perform perspective transform of the given image.
    The image can be a PIL Image or a Tensor, in which case it is expected
    to have [..., H, W] shape, where ... means an arbitrary number of leading dimensions.
565
566

    Args:
567
568
569
570
571
        img (PIL Image or Tensor): Image to be transformed.
        startpoints (list of list of ints): List containing four lists of two integers corresponding to four corners
            ``[top-left, top-right, bottom-right, bottom-left]`` of the original image.
        endpoints (list of list of ints): List containing four lists of two integers corresponding to four corners
            ``[top-left, top-right, bottom-right, bottom-left]`` of the transformed image.
572
573
574
        interpolation (InterpolationMode): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.BILINEAR``.
            If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported.
575
            For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
576
        fill (sequence or int or float, optional): Pixel fill value for the area outside the transformed
577
            image. If int or float, the value is used for all bands respectively.
578
579
580
581
            This option is supported for PIL image and Tensor inputs.
            In torchscript mode single int/float value is not supported, please use a tuple
            or list of length 1: ``[value, ]``.
            If input is PIL Image, the options is only available for ``Pillow>=5.0.0``.
582

583
    Returns:
584
        PIL Image or Tensor: transformed Image.
585
    """
586

587
    coeffs = _get_perspective_coeffs(startpoints, endpoints)
588

589
590
591
    # Backward compatibility with integer value
    if isinstance(interpolation, int):
        warnings.warn(
592
593
            "Argument interpolation should be of type InterpolationMode instead of int. "
            "Please, use InterpolationMode enum."
594
595
596
        )
        interpolation = _interpolation_modes_from_int(interpolation)

597
598
    if not isinstance(interpolation, InterpolationMode):
        raise TypeError("Argument interpolation should be a InterpolationMode")
599

600
    if not isinstance(img, torch.Tensor):
601
602
        pil_interpolation = pil_modes_mapping[interpolation]
        return F_pil.perspective(img, coeffs, interpolation=pil_interpolation, fill=fill)
603

604
    return F_t.perspective(img, coeffs, interpolation=interpolation.value, fill=fill)
605
606


607
608
def vflip(img: Tensor) -> Tensor:
    """Vertically flip the given PIL Image or torch Tensor.
609
610

    Args:
vfdev's avatar
vfdev committed
611
        img (PIL Image or Tensor): Image to be flipped. If img
612
613
614
            is a Tensor, it is expected to be in [..., H, W] format,
            where ... means it can have an arbitrary number of trailing
            dimensions.
615
616
617
618

    Returns:
        PIL Image:  Vertically flipped image.
    """
619
620
    if not isinstance(img, torch.Tensor):
        return F_pil.vflip(img)
621

622
    return F_t.vflip(img)
623
624


vfdev's avatar
vfdev committed
625
626
627
628
def five_crop(img: Tensor, size: List[int]) -> Tuple[Tensor, Tensor, Tensor, Tensor, Tensor]:
    """Crop the given image into four corners and the central crop.
    The image can be a PIL Image or a Tensor, in which case it is expected
    to have [..., H, W] shape, where ... means an arbitrary number of leading dimensions
629
630
631
632
633
634

    .. Note::
        This transform returns a tuple of images and there may be a
        mismatch in the number of inputs and targets your ``Dataset`` returns.

    Args:
vfdev's avatar
vfdev committed
635
636
637
638
        img (PIL Image or Tensor): Image to be cropped.
        size (sequence or int): Desired output size of the crop. If size is an
            int instead of sequence like (h, w), a square crop (size, size) is
            made. If provided a tuple or list of length 1, it will be interpreted as (size[0], size[0]).
639

640
    Returns:
641
642
       tuple: tuple (tl, tr, bl, br, center)
                Corresponding top left, top right, bottom left, bottom right and center crop.
643
644
645
    """
    if isinstance(size, numbers.Number):
        size = (int(size), int(size))
vfdev's avatar
vfdev committed
646
647
    elif isinstance(size, (tuple, list)) and len(size) == 1:
        size = (size[0], size[0])
648

vfdev's avatar
vfdev committed
649
650
651
652
    if len(size) != 2:
        raise ValueError("Please provide only two dimensions (h, w) for size.")

    image_width, image_height = _get_image_size(img)
653
654
655
656
657
    crop_height, crop_width = size
    if crop_width > image_width or crop_height > image_height:
        msg = "Requested crop size {} is bigger than input size {}"
        raise ValueError(msg.format(size, (image_height, image_width)))

vfdev's avatar
vfdev committed
658
659
660
661
662
663
664
665
    tl = crop(img, 0, 0, crop_height, crop_width)
    tr = crop(img, 0, image_width - crop_width, crop_height, crop_width)
    bl = crop(img, image_height - crop_height, 0, crop_height, crop_width)
    br = crop(img, image_height - crop_height, image_width - crop_width, crop_height, crop_width)

    center = center_crop(img, [crop_height, crop_width])

    return tl, tr, bl, br, center
666
667


vfdev's avatar
vfdev committed
668
669
670
def ten_crop(img: Tensor, size: List[int], vertical_flip: bool = False) -> List[Tensor]:
    """Generate ten cropped images from the given image.
    Crop the given image into four corners and the central crop plus the
671
    flipped version of these (horizontal flipping is used by default).
vfdev's avatar
vfdev committed
672
673
    The image can be a PIL Image or a Tensor, in which case it is expected
    to have [..., H, W] shape, where ... means an arbitrary number of leading dimensions
674
675
676
677
678

    .. Note::
        This transform returns a tuple of images and there may be a
        mismatch in the number of inputs and targets your ``Dataset`` returns.

679
    Args:
vfdev's avatar
vfdev committed
680
        img (PIL Image or Tensor): Image to be cropped.
681
        size (sequence or int): Desired output size of the crop. If size is an
682
            int instead of sequence like (h, w), a square crop (size, size) is
vfdev's avatar
vfdev committed
683
            made. If provided a tuple or list of length 1, it will be interpreted as (size[0], size[0]).
684
        vertical_flip (bool): Use vertical flipping instead of horizontal
685
686

    Returns:
687
688
689
        tuple: tuple (tl, tr, bl, br, center, tl_flip, tr_flip, bl_flip, br_flip, center_flip)
            Corresponding top left, top right, bottom left, bottom right and
            center crop and same for the flipped image.
690
691
692
    """
    if isinstance(size, numbers.Number):
        size = (int(size), int(size))
vfdev's avatar
vfdev committed
693
694
695
696
697
    elif isinstance(size, (tuple, list)) and len(size) == 1:
        size = (size[0], size[0])

    if len(size) != 2:
        raise ValueError("Please provide only two dimensions (h, w) for size.")
698
699
700
701
702
703
704
705
706
707
708
709

    first_five = five_crop(img, size)

    if vertical_flip:
        img = vflip(img)
    else:
        img = hflip(img)

    second_five = five_crop(img, size)
    return first_five + second_five


710
def adjust_brightness(img: Tensor, brightness_factor: float) -> Tensor:
711
712
713
    """Adjust brightness of an Image.

    Args:
vfdev's avatar
vfdev committed
714
        img (PIL Image or Tensor): Image to be adjusted.
715
716
717
718
719
        brightness_factor (float):  How much to adjust the brightness. Can be
            any non negative number. 0 gives a black image, 1 gives the
            original image while 2 increases the brightness by a factor of 2.

    Returns:
vfdev's avatar
vfdev committed
720
        PIL Image or Tensor: Brightness adjusted image.
721
    """
722
723
    if not isinstance(img, torch.Tensor):
        return F_pil.adjust_brightness(img, brightness_factor)
724

725
    return F_t.adjust_brightness(img, brightness_factor)
726
727


728
def adjust_contrast(img: Tensor, contrast_factor: float) -> Tensor:
729
730
731
    """Adjust contrast of an Image.

    Args:
vfdev's avatar
vfdev committed
732
        img (PIL Image or Tensor): Image to be adjusted.
733
734
735
736
737
        contrast_factor (float): How much to adjust the contrast. Can be any
            non negative number. 0 gives a solid gray image, 1 gives the
            original image while 2 increases the contrast by a factor of 2.

    Returns:
vfdev's avatar
vfdev committed
738
        PIL Image or Tensor: Contrast adjusted image.
739
    """
740
741
    if not isinstance(img, torch.Tensor):
        return F_pil.adjust_contrast(img, contrast_factor)
742

743
    return F_t.adjust_contrast(img, contrast_factor)
744
745


746
def adjust_saturation(img: Tensor, saturation_factor: float) -> Tensor:
747
748
749
    """Adjust color saturation of an image.

    Args:
vfdev's avatar
vfdev committed
750
        img (PIL Image or Tensor): Image to be adjusted.
751
752
753
754
755
        saturation_factor (float):  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.

    Returns:
vfdev's avatar
vfdev committed
756
        PIL Image or Tensor: Saturation adjusted image.
757
    """
758
759
    if not isinstance(img, torch.Tensor):
        return F_pil.adjust_saturation(img, saturation_factor)
760

761
    return F_t.adjust_saturation(img, saturation_factor)
762
763


764
def adjust_hue(img: Tensor, hue_factor: float) -> Tensor:
765
766
767
768
769
770
771
772
773
    """Adjust hue of an image.

    The image hue is adjusted by converting the image to HSV and
    cyclically shifting the intensities in the hue channel (H).
    The image is then converted back to original image mode.

    `hue_factor` is the amount of shift in H channel and must be in the
    interval `[-0.5, 0.5]`.

774
775
776
    See `Hue`_ for more details.

    .. _Hue: https://en.wikipedia.org/wiki/Hue
777
778

    Args:
779
        img (PIL Image or Tensor): Image to be adjusted.
780
781
782
783
784
785
786
        hue_factor (float):  How much to shift the hue channel. Should be in
            [-0.5, 0.5]. 0.5 and -0.5 give complete reversal of hue channel in
            HSV space in positive and negative direction respectively.
            0 means no shift. Therefore, both -0.5 and 0.5 will give an image
            with complementary colors while 0 gives the original image.

    Returns:
787
        PIL Image or Tensor: Hue adjusted image.
788
    """
789
790
    if not isinstance(img, torch.Tensor):
        return F_pil.adjust_hue(img, hue_factor)
791

792
    return F_t.adjust_hue(img, hue_factor)
793
794


795
def adjust_gamma(img: Tensor, gamma: float, gain: float = 1) -> Tensor:
796
    r"""Perform gamma correction on an image.
797
798
799
800

    Also known as Power Law Transform. Intensities in RGB mode are adjusted
    based on the following equation:

801
802
803
804
    .. math::
        I_{\text{out}} = 255 \times \text{gain} \times \left(\frac{I_{\text{in}}}{255}\right)^{\gamma}

    See `Gamma Correction`_ for more details.
805

806
    .. _Gamma Correction: https://en.wikipedia.org/wiki/Gamma_correction
807
808

    Args:
809
        img (PIL Image or Tensor): PIL Image to be adjusted.
810
811
812
        gamma (float): Non negative real number, same as :math:`\gamma` in the equation.
            gamma larger than 1 make the shadows darker,
            while gamma smaller than 1 make dark regions lighter.
813
        gain (float): The constant multiplier.
814
815
    Returns:
        PIL Image or Tensor: Gamma correction adjusted image.
816
    """
817
818
    if not isinstance(img, torch.Tensor):
        return F_pil.adjust_gamma(img, gamma, gain)
819

820
    return F_t.adjust_gamma(img, gamma, gain)
821
822


vfdev's avatar
vfdev committed
823
def _get_inverse_affine_matrix(
vfdev's avatar
vfdev committed
824
        center: List[float], angle: float, translate: List[float], scale: float, shear: List[float]
vfdev's avatar
vfdev committed
825
) -> List[float]:
826
827
828
829
830
831
832
    # Helper method to compute inverse matrix for affine transformation

    # As it is explained in PIL.Image.rotate
    # We need compute INVERSE of affine transformation matrix: M = T * C * RSS * C^-1
    # where T is translation matrix: [1, 0, tx | 0, 1, ty | 0, 0, 1]
    #       C is translation matrix to keep center: [1, 0, cx | 0, 1, cy | 0, 0, 1]
    #       RSS is rotation with scale and shear matrix
833
834
835
836
837
838
839
840
841
842
    #       RSS(a, s, (sx, sy)) =
    #       = R(a) * S(s) * SHy(sy) * SHx(sx)
    #       = [ s*cos(a - sy)/cos(sy), s*(-cos(a - sy)*tan(x)/cos(y) - sin(a)), 0 ]
    #         [ s*sin(a + sy)/cos(sy), s*(-sin(a - sy)*tan(x)/cos(y) + cos(a)), 0 ]
    #         [ 0                    , 0                                      , 1 ]
    #
    # where R is a rotation matrix, S is a scaling matrix, and SHx and SHy are the shears:
    # SHx(s) = [1, -tan(s)] and SHy(s) = [1      , 0]
    #          [0, 1      ]              [-tan(s), 1]
    #
843
844
    # Thus, the inverse is M^-1 = C * RSS^-1 * C^-1 * T^-1

845
846
847
848
849
850
851
    rot = math.radians(angle)
    sx, sy = [math.radians(s) for s in shear]

    cx, cy = center
    tx, ty = translate

    # RSS without scaling
vfdev's avatar
vfdev committed
852
853
854
855
    a = math.cos(rot - sy) / math.cos(sy)
    b = -math.cos(rot - sy) * math.tan(sx) / math.cos(sy) - math.sin(rot)
    c = math.sin(rot - sy) / math.cos(sy)
    d = -math.sin(rot - sy) * math.tan(sx) / math.cos(sy) + math.cos(rot)
856
857

    # Inverted rotation matrix with scale and shear
858
    # det([[a, b], [c, d]]) == 1, since det(rotation) = 1 and det(shear) = 1
vfdev's avatar
vfdev committed
859
860
    matrix = [d, -b, 0.0, -c, a, 0.0]
    matrix = [x / scale for x in matrix]
861
862

    # Apply inverse of translation and of center translation: RSS^-1 * C^-1 * T^-1
vfdev's avatar
vfdev committed
863
864
    matrix[2] += matrix[0] * (-cx - tx) + matrix[1] * (-cy - ty)
    matrix[5] += matrix[3] * (-cx - tx) + matrix[4] * (-cy - ty)
865
866

    # Apply center translation: C * RSS^-1 * C^-1 * T^-1
vfdev's avatar
vfdev committed
867
868
    matrix[2] += cx
    matrix[5] += cy
869

vfdev's avatar
vfdev committed
870
    return matrix
871

vfdev's avatar
vfdev committed
872

vfdev's avatar
vfdev committed
873
def rotate(
874
        img: Tensor, angle: float, interpolation: InterpolationMode = InterpolationMode.NEAREST,
875
        expand: bool = False, center: Optional[List[int]] = None,
876
        fill: Optional[List[float]] = None, resample: Optional[int] = None
vfdev's avatar
vfdev committed
877
878
879
880
881
882
883
884
) -> Tensor:
    """Rotate the image by angle.
    The image can be a PIL Image or a Tensor, in which case it is expected
    to have [..., H, W] shape, where ... means an arbitrary number of leading dimensions.

    Args:
        img (PIL Image or Tensor): image to be rotated.
        angle (float or int): rotation angle value in degrees, counter-clockwise.
885
886
887
        interpolation (InterpolationMode): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.NEAREST``.
            If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported.
888
            For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
vfdev's avatar
vfdev committed
889
890
891
892
893
894
        expand (bool, optional): Optional expansion flag.
            If true, expands the output image to make it large enough to hold the entire rotated image.
            If false or omitted, make the output image the same size as the input image.
            Note that the expand flag assumes rotation around the center and no translation.
        center (list or tuple, optional): Optional center of rotation. Origin is the upper left corner.
            Default is the center of the image.
895
        fill (sequence or int or float, optional): Pixel fill value for the area outside the transformed
vfdev's avatar
vfdev committed
896
            image. If int or float, the value is used for all bands respectively.
897
898
899
900
            This option is supported for PIL image and Tensor inputs.
            In torchscript mode single int/float value is not supported, please use a tuple
            or list of length 1: ``[value, ]``.
            If input is PIL Image, the options is only available for ``Pillow>=5.2.0``.
vfdev's avatar
vfdev committed
901
902
903
904
905
906
907

    Returns:
        PIL Image or Tensor: Rotated image.

    .. _filters: https://pillow.readthedocs.io/en/latest/handbook/concepts.html#filters

    """
908
909
910
911
912
913
914
915
916
    if resample is not None:
        warnings.warn(
            "Argument resample is deprecated and will be removed since v0.10.0. Please, use interpolation instead"
        )
        interpolation = _interpolation_modes_from_int(resample)

    # Backward compatibility with integer value
    if isinstance(interpolation, int):
        warnings.warn(
917
918
            "Argument interpolation should be of type InterpolationMode instead of int. "
            "Please, use InterpolationMode enum."
919
920
921
        )
        interpolation = _interpolation_modes_from_int(interpolation)

vfdev's avatar
vfdev committed
922
923
924
925
926
927
    if not isinstance(angle, (int, float)):
        raise TypeError("Argument angle should be int or float")

    if center is not None and not isinstance(center, (list, tuple)):
        raise TypeError("Argument center should be a sequence")

928
929
    if not isinstance(interpolation, InterpolationMode):
        raise TypeError("Argument interpolation should be a InterpolationMode")
930

vfdev's avatar
vfdev committed
931
    if not isinstance(img, torch.Tensor):
932
933
        pil_interpolation = pil_modes_mapping[interpolation]
        return F_pil.rotate(img, angle=angle, interpolation=pil_interpolation, expand=expand, center=center, fill=fill)
vfdev's avatar
vfdev committed
934
935
936
937

    center_f = [0.0, 0.0]
    if center is not None:
        img_size = _get_image_size(img)
938
939
940
        # Center values should be in pixel coordinates but translated such that (0, 0) corresponds to image center.
        center_f = [1.0 * (c - s * 0.5) for c, s in zip(center, img_size)]

vfdev's avatar
vfdev committed
941
942
943
    # due to current incoherence of rotation angle direction between affine and rotate implementations
    # we need to set -angle.
    matrix = _get_inverse_affine_matrix(center_f, -angle, [0.0, 0.0], 1.0, [0.0, 0.0])
944
    return F_t.rotate(img, matrix=matrix, interpolation=interpolation.value, expand=expand, fill=fill)
vfdev's avatar
vfdev committed
945
946


vfdev's avatar
vfdev committed
947
948
def affine(
        img: Tensor, angle: float, translate: List[int], scale: float, shear: List[float],
949
950
        interpolation: InterpolationMode = InterpolationMode.NEAREST, fill: Optional[List[float]] = None,
        resample: Optional[int] = None, fillcolor: Optional[List[float]] = None
vfdev's avatar
vfdev committed
951
952
953
954
) -> Tensor:
    """Apply affine transformation on the image keeping image center invariant.
    The image can be a PIL Image or a Tensor, in which case it is expected
    to have [..., H, W] shape, where ... means an arbitrary number of leading dimensions.
955
956

    Args:
vfdev's avatar
vfdev committed
957
        img (PIL Image or Tensor): image to transform.
958
        angle (float or int): rotation angle in degrees between -180 and 180, clockwise direction.
959
960
        translate (list or tuple of integers): horizontal and vertical translations (post-rotation translation)
        scale (float): overall scale
ptrblck's avatar
ptrblck committed
961
        shear (float or tuple or list): shear angle value in degrees between -180 to 180, clockwise direction.
vfdev's avatar
vfdev committed
962
963
            If a tuple of list is specified, the first value corresponds to a shear parallel to the x axis, while
            the second value corresponds to a shear parallel to the y axis.
964
965
966
        interpolation (InterpolationMode): Desired interpolation enum defined by
            :class:`torchvision.transforms.InterpolationMode`. Default is ``InterpolationMode.NEAREST``.
            If input is Tensor, only ``InterpolationMode.NEAREST``, ``InterpolationMode.BILINEAR`` are supported.
967
            For backward compatibility integer values (e.g. ``PIL.Image.NEAREST``) are still acceptable.
968
969
970
971
972
973
974
        fill (sequence or int or float, optional): Pixel fill value for the area outside the transformed
            image. If int or float, the value is used for all bands respectively.
            This option is supported for PIL image and Tensor inputs.
            In torchscript mode single int/float value is not supported, please use a tuple
            or list of length 1: ``[value, ]``.
            If input is PIL Image, the options is only available for ``Pillow>=5.0.0``.
        fillcolor (sequence, int, float): deprecated argument and will be removed since v0.10.0.
975
976
977
            Please use `arg`:fill: instead.
        resample (int, optional): deprecated argument and will be removed since v0.10.0.
            Please use `arg`:interpolation: instead.
vfdev's avatar
vfdev committed
978
979
980

    Returns:
        PIL Image or Tensor: Transformed image.
981
    """
982
983
984
985
986
987
988
989
990
    if resample is not None:
        warnings.warn(
            "Argument resample is deprecated and will be removed since v0.10.0. Please, use interpolation instead"
        )
        interpolation = _interpolation_modes_from_int(resample)

    # Backward compatibility with integer value
    if isinstance(interpolation, int):
        warnings.warn(
991
992
            "Argument interpolation should be of type InterpolationMode instead of int. "
            "Please, use InterpolationMode enum."
993
994
995
996
997
998
999
1000
1001
        )
        interpolation = _interpolation_modes_from_int(interpolation)

    if fillcolor is not None:
        warnings.warn(
            "Argument fillcolor is deprecated and will be removed since v0.10.0. Please, use fill instead"
        )
        fill = fillcolor

vfdev's avatar
vfdev committed
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
    if not isinstance(angle, (int, float)):
        raise TypeError("Argument angle should be int or float")

    if not isinstance(translate, (list, tuple)):
        raise TypeError("Argument translate should be a sequence")

    if len(translate) != 2:
        raise ValueError("Argument translate should be a sequence of length 2")

    if scale <= 0.0:
        raise ValueError("Argument scale should be positive")

    if not isinstance(shear, (numbers.Number, (list, tuple))):
        raise TypeError("Shear should be either a single value or a sequence of two values")

1017
1018
    if not isinstance(interpolation, InterpolationMode):
        raise TypeError("Argument interpolation should be a InterpolationMode")
1019

vfdev's avatar
vfdev committed
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
    if isinstance(angle, int):
        angle = float(angle)

    if isinstance(translate, tuple):
        translate = list(translate)

    if isinstance(shear, numbers.Number):
        shear = [shear, 0.0]

    if isinstance(shear, tuple):
        shear = list(shear)

    if len(shear) == 1:
        shear = [shear[0], shear[0]]

    if len(shear) != 2:
        raise ValueError("Shear should be a sequence containing two values. Got {}".format(shear))

    img_size = _get_image_size(img)
    if not isinstance(img, torch.Tensor):
        # center = (img_size[0] * 0.5 + 0.5, img_size[1] * 0.5 + 0.5)
        # it is visually better to estimate the center without 0.5 offset
        # otherwise image rotated by 90 degrees is shifted vs output image of torch.rot90 or F_t.affine
        center = [img_size[0] * 0.5, img_size[1] * 0.5]
        matrix = _get_inverse_affine_matrix(center, angle, translate, scale, shear)
1045
1046
        pil_interpolation = pil_modes_mapping[interpolation]
        return F_pil.affine(img, matrix=matrix, interpolation=pil_interpolation, fill=fill)
1047

1048
1049
    translate_f = [1.0 * t for t in translate]
    matrix = _get_inverse_affine_matrix([0.0, 0.0], angle, translate_f, scale, shear)
1050
    return F_t.affine(img, matrix=matrix, interpolation=interpolation.value, fill=fill)
1051
1052


1053
@torch.jit.unused
1054
def to_grayscale(img, num_output_channels=1):
1055
    """Convert PIL image of any mode (RGB, HSV, LAB, etc) to grayscale version of image.
1056
1057

    Args:
1058
1059
        img (PIL Image): PIL Image to be converted to grayscale.
        num_output_channels (int): number of channels of the output image. Value can be 1 or 3. Default, 1.
1060
1061

    Returns:
1062
1063
1064
1065
        PIL Image: Grayscale version of the image.
            if num_output_channels = 1 : returned image is single channel

            if num_output_channels = 3 : returned image is 3 channel with r = g = b
1066
    """
1067
1068
    if isinstance(img, Image.Image):
        return F_pil.to_grayscale(img, num_output_channels)
1069

1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
    raise TypeError("Input should be PIL Image")


def rgb_to_grayscale(img: Tensor, num_output_channels: int = 1) -> Tensor:
    """Convert RGB image to grayscale version of image.
    The image can be a PIL Image or a Tensor, in which case it is expected
    to have [..., H, W] shape, where ... means an arbitrary number of leading dimensions

    Note:
        Please, note that this method supports only RGB images as input. For inputs in other color spaces,
        please, consider using meth:`~torchvision.transforms.functional.to_grayscale` with PIL Image.

    Args:
        img (PIL Image or Tensor): RGB Image to be converted to grayscale.
        num_output_channels (int): number of channels of the output image. Value can be 1 or 3. Default, 1.

    Returns:
        PIL Image or Tensor: Grayscale version of the image.
            if num_output_channels = 1 : returned image is single channel

            if num_output_channels = 3 : returned image is 3 channel with r = g = b
    """
    if not isinstance(img, torch.Tensor):
        return F_pil.to_grayscale(img, num_output_channels)

    return F_t.rgb_to_grayscale(img, num_output_channels)
1096
1097


1098
def erase(img: Tensor, i: int, j: int, h: int, w: int, v: Tensor, inplace: bool = False) -> Tensor:
1099
1100
1101
1102
1103
1104
1105
1106
1107
    """ Erase the input Tensor Image with given value.

    Args:
        img (Tensor Image): Tensor image of size (C, H, W) to be erased
        i (int): i in (i,j) i.e coordinates of the upper left corner.
        j (int): j in (i,j) i.e coordinates of the upper left corner.
        h (int): Height of the erased region.
        w (int): Width of the erased region.
        v: Erasing value.
Zhun Zhong's avatar
Zhun Zhong committed
1108
        inplace(bool, optional): For in-place operations. By default is set False.
1109
1110
1111
1112
1113
1114
1115

    Returns:
        Tensor Image: Erased image.
    """
    if not isinstance(img, torch.Tensor):
        raise TypeError('img should be Tensor Image. Got {}'.format(type(img)))

1116
1117
1118
    if not inplace:
        img = img.clone()

vfdev's avatar
vfdev committed
1119
    img[..., i:i + h, j:j + w] = v
1120
    return img
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180


def gaussian_blur(img: Tensor, kernel_size: List[int], sigma: Optional[List[float]] = None) -> Tensor:
    """Performs Gaussian blurring on the img by given kernel.
    The image can be a PIL Image or a Tensor, in which case it is expected
    to have [..., H, W] shape, where ... means an arbitrary number of leading dimensions

    Args:
        img (PIL Image or Tensor): Image to be blurred
        kernel_size (sequence of ints or int): Gaussian kernel size. Can be a sequence of integers
            like ``(kx, ky)`` or a single integer for square kernels.
            In torchscript mode kernel_size as single int is not supported, use a tuple or
            list of length 1: ``[ksize, ]``.
        sigma (sequence of floats or float, optional): Gaussian kernel standard deviation. Can be a
            sequence of floats like ``(sigma_x, sigma_y)`` or a single float to define the
            same sigma in both X/Y directions. If None, then it is computed using
            ``kernel_size`` as ``sigma = 0.3 * ((kernel_size - 1) * 0.5 - 1) + 0.8``.
            Default, None. In torchscript mode sigma as single float is
            not supported, use a tuple or list of length 1: ``[sigma, ]``.

    Returns:
        PIL Image or Tensor: Gaussian Blurred version of the image.
    """
    if not isinstance(kernel_size, (int, list, tuple)):
        raise TypeError('kernel_size should be int or a sequence of integers. Got {}'.format(type(kernel_size)))
    if isinstance(kernel_size, int):
        kernel_size = [kernel_size, kernel_size]
    if len(kernel_size) != 2:
        raise ValueError('If kernel_size is a sequence its length should be 2. Got {}'.format(len(kernel_size)))
    for ksize in kernel_size:
        if ksize % 2 == 0 or ksize < 0:
            raise ValueError('kernel_size should have odd and positive integers. Got {}'.format(kernel_size))

    if sigma is None:
        sigma = [ksize * 0.15 + 0.35 for ksize in kernel_size]

    if sigma is not None and not isinstance(sigma, (int, float, list, tuple)):
        raise TypeError('sigma should be either float or sequence of floats. Got {}'.format(type(sigma)))
    if isinstance(sigma, (int, float)):
        sigma = [float(sigma), float(sigma)]
    if isinstance(sigma, (list, tuple)) and len(sigma) == 1:
        sigma = [sigma[0], sigma[0]]
    if len(sigma) != 2:
        raise ValueError('If sigma is a sequence, its length should be 2. Got {}'.format(len(sigma)))
    for s in sigma:
        if s <= 0.:
            raise ValueError('sigma should have positive values. Got {}'.format(sigma))

    t_img = img
    if not isinstance(img, torch.Tensor):
        if not F_pil._is_pil_image(img):
            raise TypeError('img should be PIL Image or Tensor. Got {}'.format(type(img)))

        t_img = to_tensor(img)

    output = F_t.gaussian_blur(t_img, kernel_size, sigma)

    if not isinstance(img, torch.Tensor):
        output = to_pil_image(output)
    return output