File size: 22,680 Bytes
24f9881
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
# MIT License

# Copyright (c) 2022 Intelligent Systems Lab Org

# Permission is hereby granted, free of charge, to any person obtaining a copy
# of this software and associated documentation files (the "Software"), to deal
# in the Software without restriction, including without limitation the rights
# to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
# copies of the Software, and to permit persons to whom the Software is
# furnished to do so, subject to the following conditions:

# The above copyright notice and this permission notice shall be included in all
# copies or substantial portions of the Software.

# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
# AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
# OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
# SOFTWARE.

# File author: Shariq Farooq Bhat

# This file is partly inspired from BTS (https://github.com/cleinc/bts/blob/master/pytorch/bts_dataloader.py); author: Jin Han Lee

import itertools
import os
import random

import numpy as np
import cv2
import torch
import torch.nn as nn
import torch.utils.data.distributed
from zoedepth.utils.easydict import EasyDict as edict
from PIL import Image, ImageOps
from torch.utils.data import DataLoader, Dataset
from torchvision import transforms

from zoedepth.utils.config import change_dataset

from .ddad import get_ddad_loader
from .diml_indoor_test import get_diml_indoor_loader
from .diml_outdoor_test import get_diml_outdoor_loader
from .diode import get_diode_loader
from .hypersim import get_hypersim_loader
from .ibims import get_ibims_loader
from .sun_rgbd_loader import get_sunrgbd_loader
from .vkitti import get_vkitti_loader
from .vkitti2 import get_vkitti2_loader

from .preprocess import CropParams, get_white_border, get_black_border


def _is_pil_image(img):
    return isinstance(img, Image.Image)


def _is_numpy_image(img):
    return isinstance(img, np.ndarray) and (img.ndim in {2, 3})


def preprocessing_transforms(mode, **kwargs):
    return transforms.Compose([
        ToTensor(mode=mode, **kwargs)
    ])


class DepthDataLoader(object):
    def __init__(self, config, mode, device='cpu', transform=None, **kwargs):
        """
        Data loader for depth datasets

        Args:
            config (dict): Config dictionary. Refer to utils/config.py
            mode (str): "train" or "online_eval"
            device (str, optional): Device to load the data on. Defaults to 'cpu'.
            transform (torchvision.transforms, optional): Transform to apply to the data. Defaults to None.
        """

        self.config = config

        if config.dataset == 'ibims':
            self.data = get_ibims_loader(config, batch_size=1, num_workers=1)
            return

        if config.dataset == 'sunrgbd':
            self.data = get_sunrgbd_loader(
                data_dir_root=config.sunrgbd_root, batch_size=1, num_workers=1)
            return

        if config.dataset == 'diml_indoor':
            self.data = get_diml_indoor_loader(
                data_dir_root=config.diml_indoor_root, batch_size=1, num_workers=1)
            return

        if config.dataset == 'diml_outdoor':
            self.data = get_diml_outdoor_loader(
                data_dir_root=config.diml_outdoor_root, batch_size=1, num_workers=1)
            return

        if "diode" in config.dataset:
            self.data = get_diode_loader(
                config[config.dataset+"_root"], batch_size=1, num_workers=1)
            return

        if config.dataset == 'hypersim_test':
            self.data = get_hypersim_loader(
                config.hypersim_test_root, batch_size=1, num_workers=1)
            return

        if config.dataset == 'vkitti':
            self.data = get_vkitti_loader(
                config.vkitti_root, batch_size=1, num_workers=1)
            return

        if config.dataset == 'vkitti2':
            self.data = get_vkitti2_loader(
                config.vkitti2_root, batch_size=1, num_workers=1)
            return

        if config.dataset == 'ddad':
            self.data = get_ddad_loader(config.ddad_root, resize_shape=(
                352, 1216), batch_size=1, num_workers=1)
            return

        img_size = self.config.get("img_size", None)
        img_size = img_size if self.config.get(
            "do_input_resize", False) else None

        if transform is None:
            transform = preprocessing_transforms(mode, size=img_size)

        if mode == 'train':

            Dataset = DataLoadPreprocess
            self.training_samples = Dataset(
                config, mode, transform=transform, device=device)

            if config.distributed:
                self.train_sampler = torch.utils.data.distributed.DistributedSampler(
                    self.training_samples)
            else:
                self.train_sampler = None

            self.data = DataLoader(self.training_samples,
                                   batch_size=config.batch_size,
                                   shuffle=(self.train_sampler is None),
                                   num_workers=config.workers,
                                   pin_memory=True,
                                   persistent_workers=True,
                                #    prefetch_factor=2,
                                   sampler=self.train_sampler)

        elif mode == 'online_eval':
            self.testing_samples = DataLoadPreprocess(
                config, mode, transform=transform)
            if config.distributed:  # redundant. here only for readability and to be more explicit
                # Give whole test set to all processes (and report evaluation only on one) regardless
                self.eval_sampler = None
            else:
                self.eval_sampler = None
            self.data = DataLoader(self.testing_samples, 1,
                                   shuffle=kwargs.get("shuffle_test", False),
                                   num_workers=1,
                                   pin_memory=False,
                                   sampler=self.eval_sampler)

        elif mode == 'test':
            self.testing_samples = DataLoadPreprocess(
                config, mode, transform=transform)
            self.data = DataLoader(self.testing_samples,
                                   1, shuffle=False, num_workers=1)

        else:
            print(
                'mode should be one of \'train, test, online_eval\'. Got {}'.format(mode))


def repetitive_roundrobin(*iterables):
    """
    cycles through iterables but sample wise
    first yield first sample from first iterable then first sample from second iterable and so on
    then second sample from first iterable then second sample from second iterable and so on

    If one iterable is shorter than the others, it is repeated until all iterables are exhausted
    repetitive_roundrobin('ABC', 'D', 'EF') --> A D E B D F C D E
    """
    # Repetitive roundrobin
    iterables_ = [iter(it) for it in iterables]
    exhausted = [False] * len(iterables)
    while not all(exhausted):
        for i, it in enumerate(iterables_):
            try:
                yield next(it)
            except StopIteration:
                exhausted[i] = True
                iterables_[i] = itertools.cycle(iterables[i])
                # First elements may get repeated if one iterable is shorter than the others
                yield next(iterables_[i])


class RepetitiveRoundRobinDataLoader(object):
    def __init__(self, *dataloaders):
        self.dataloaders = dataloaders

    def __iter__(self):
        return repetitive_roundrobin(*self.dataloaders)

    def __len__(self):
        # First samples get repeated, thats why the plus one
        return len(self.dataloaders) * (max(len(dl) for dl in self.dataloaders) + 1)


class MixedNYUKITTI(object):
    def __init__(self, config, mode, device='cpu', **kwargs):
        config = edict(config)
        config.workers = config.workers // 2
        self.config = config
        nyu_conf = change_dataset(edict(config), 'nyu')
        kitti_conf = change_dataset(edict(config), 'kitti')

        # make nyu default for testing
        self.config = config = nyu_conf
        img_size = self.config.get("img_size", None)
        img_size = img_size if self.config.get(
            "do_input_resize", False) else None
        if mode == 'train':
            nyu_loader = DepthDataLoader(
                nyu_conf, mode, device=device, transform=preprocessing_transforms(mode, size=img_size)).data
            kitti_loader = DepthDataLoader(
                kitti_conf, mode, device=device, transform=preprocessing_transforms(mode, size=img_size)).data
            # It has been changed to repetitive roundrobin
            self.data = RepetitiveRoundRobinDataLoader(
                nyu_loader, kitti_loader)
        else:
            self.data = DepthDataLoader(nyu_conf, mode, device=device).data


def remove_leading_slash(s):
    if s[0] == '/' or s[0] == '\\':
        return s[1:]
    return s


class CachedReader:
    def __init__(self, shared_dict=None):
        if shared_dict:
            self._cache = shared_dict
        else:
            self._cache = {}

    def open(self, fpath):
        im = self._cache.get(fpath, None)
        if im is None:
            im = self._cache[fpath] = Image.open(fpath)
        return im


class ImReader:
    def __init__(self):
        pass

    # @cache
    def open(self, fpath):
        return Image.open(fpath)


class DataLoadPreprocess(Dataset):
    def __init__(self, config, mode, transform=None, is_for_online_eval=False, **kwargs):
        self.config = config
        if mode == 'online_eval':
            with open(config.filenames_file_eval, 'r') as f:
                self.filenames = f.readlines()
        else:
            with open(config.filenames_file, 'r') as f:
                self.filenames = f.readlines()

        self.mode = mode
        self.transform = transform
        self.to_tensor = ToTensor(mode)
        self.is_for_online_eval = is_for_online_eval
        if config.use_shared_dict:
            self.reader = CachedReader(config.shared_dict)
        else:
            self.reader = ImReader()

    def postprocess(self, sample):
        return sample

    def __getitem__(self, idx):
        sample_path = self.filenames[idx]
        focal = float(sample_path.split()[2])
        sample = {}

        if self.mode == 'train':
            if self.config.dataset == 'kitti' and self.config.use_right and random.random() > 0.5:
                image_path = os.path.join(
                    self.config.data_path, remove_leading_slash(sample_path.split()[3]))
                depth_path = os.path.join(
                    self.config.gt_path, remove_leading_slash(sample_path.split()[4]))
            else:
                image_path = os.path.join(
                    self.config.data_path, remove_leading_slash(sample_path.split()[0]))
                depth_path = os.path.join(
                    self.config.gt_path, remove_leading_slash(sample_path.split()[1]))

            image = self.reader.open(image_path)
            depth_gt = self.reader.open(depth_path)
            w, h = image.size

            if self.config.do_kb_crop:
                height = image.height
                width = image.width
                top_margin = int(height - 352)
                left_margin = int((width - 1216) / 2)
                depth_gt = depth_gt.crop(
                    (left_margin, top_margin, left_margin + 1216, top_margin + 352))
                image = image.crop(
                    (left_margin, top_margin, left_margin + 1216, top_margin + 352))

            # Avoid blank boundaries due to pixel registration?
            # Train images have white border. Test images have black border.
            if self.config.dataset == 'nyu' and self.config.avoid_boundary:
                # print("Avoiding Blank Boundaries!")
                # We just crop and pad again with reflect padding to original size
                # original_size = image.size
                crop_params = get_white_border(np.array(image, dtype=np.uint8))
                image = image.crop((crop_params.left, crop_params.top, crop_params.right, crop_params.bottom))
                depth_gt = depth_gt.crop((crop_params.left, crop_params.top, crop_params.right, crop_params.bottom))

                # Use reflect padding to fill the blank
                image = np.array(image)
                image = np.pad(image, ((crop_params.top, h - crop_params.bottom), (crop_params.left, w - crop_params.right), (0, 0)), mode='reflect')
                image = Image.fromarray(image)

                depth_gt = np.array(depth_gt)
                depth_gt = np.pad(depth_gt, ((crop_params.top, h - crop_params.bottom), (crop_params.left, w - crop_params.right)), 'constant', constant_values=0)
                depth_gt = Image.fromarray(depth_gt)


            if self.config.do_random_rotate and (self.config.aug):
                random_angle = (random.random() - 0.5) * 2 * self.config.degree
                image = self.rotate_image(image, random_angle)
                depth_gt = self.rotate_image(
                    depth_gt, random_angle, flag=Image.NEAREST)

            image = np.asarray(image, dtype=np.float32) / 255.0
            depth_gt = np.asarray(depth_gt, dtype=np.float32)
            depth_gt = np.expand_dims(depth_gt, axis=2)

            if self.config.dataset == 'nyu':
                depth_gt = depth_gt / 1000.0
            else:
                depth_gt = depth_gt / 256.0

            if self.config.aug and (self.config.random_crop):
                image, depth_gt = self.random_crop(
                    image, depth_gt, self.config.input_height, self.config.input_width)
            
            if self.config.aug and self.config.random_translate:
                # print("Random Translation!")
                image, depth_gt = self.random_translate(image, depth_gt, self.config.max_translation)

            image, depth_gt = self.train_preprocess(image, depth_gt)
            mask = np.logical_and(depth_gt > self.config.min_depth,
                                  depth_gt < self.config.max_depth).squeeze()[None, ...]
            sample = {'image': image, 'depth': depth_gt, 'focal': focal,
                      'mask': mask, **sample}

        else:
            if self.mode == 'online_eval':
                data_path = self.config.data_path_eval
            else:
                data_path = self.config.data_path

            image_path = os.path.join(
                data_path, remove_leading_slash(sample_path.split()[0]))
            image = np.asarray(self.reader.open(image_path),
                               dtype=np.float32) / 255.0

            if self.mode == 'online_eval':
                gt_path = self.config.gt_path_eval
                depth_path = os.path.join(
                    gt_path, remove_leading_slash(sample_path.split()[1]))
                has_valid_depth = False
                try:
                    depth_gt = self.reader.open(depth_path)
                    has_valid_depth = True
                except IOError:
                    depth_gt = False
                    # print('Missing gt for {}'.format(image_path))

                if has_valid_depth:
                    depth_gt = np.asarray(depth_gt, dtype=np.float32)
                    depth_gt = np.expand_dims(depth_gt, axis=2)
                    if self.config.dataset == 'nyu':
                        depth_gt = depth_gt / 1000.0
                    else:
                        depth_gt = depth_gt / 256.0

                    mask = np.logical_and(
                        depth_gt >= self.config.min_depth, depth_gt <= self.config.max_depth).squeeze()[None, ...]
                else:
                    mask = False

            if self.config.do_kb_crop:
                height = image.shape[0]
                width = image.shape[1]
                top_margin = int(height - 352)
                left_margin = int((width - 1216) / 2)
                image = image[top_margin:top_margin + 352,
                              left_margin:left_margin + 1216, :]
                if self.mode == 'online_eval' and has_valid_depth:
                    depth_gt = depth_gt[top_margin:top_margin +
                                        352, left_margin:left_margin + 1216, :]

            if self.mode == 'online_eval':
                sample = {'image': image, 'depth': depth_gt, 'focal': focal, 'has_valid_depth': has_valid_depth,
                          'image_path': sample_path.split()[0], 'depth_path': sample_path.split()[1],
                          'mask': mask}
            else:
                sample = {'image': image, 'focal': focal}

        if (self.mode == 'train') or ('has_valid_depth' in sample and sample['has_valid_depth']):
            mask = np.logical_and(depth_gt > self.config.min_depth,
                                  depth_gt < self.config.max_depth).squeeze()[None, ...]
            sample['mask'] = mask

        if self.transform:
            sample = self.transform(sample)

        sample = self.postprocess(sample)
        sample['dataset'] = self.config.dataset
        sample = {**sample, 'image_path': sample_path.split()[0], 'depth_path': sample_path.split()[1]}

        return sample

    def rotate_image(self, image, angle, flag=Image.BILINEAR):
        result = image.rotate(angle, resample=flag)
        return result

    def random_crop(self, img, depth, height, width):
        assert img.shape[0] >= height
        assert img.shape[1] >= width
        assert img.shape[0] == depth.shape[0]
        assert img.shape[1] == depth.shape[1]
        x = random.randint(0, img.shape[1] - width)
        y = random.randint(0, img.shape[0] - height)
        img = img[y:y + height, x:x + width, :]
        depth = depth[y:y + height, x:x + width, :]

        return img, depth
    
    def random_translate(self, img, depth, max_t=20):
        assert img.shape[0] == depth.shape[0]
        assert img.shape[1] == depth.shape[1]
        p = self.config.translate_prob
        do_translate = random.random()
        if do_translate > p:
            return img, depth
        x = random.randint(-max_t, max_t)
        y = random.randint(-max_t, max_t)
        M = np.float32([[1, 0, x], [0, 1, y]])
        # print(img.shape, depth.shape)
        img = cv2.warpAffine(img, M, (img.shape[1], img.shape[0]))
        depth = cv2.warpAffine(depth, M, (depth.shape[1], depth.shape[0]))
        depth = depth.squeeze()[..., None]  # add channel dim back. Affine warp removes it
        # print("after", img.shape, depth.shape)
        return img, depth

    def train_preprocess(self, image, depth_gt):
        if self.config.aug:
            # Random flipping
            do_flip = random.random()
            if do_flip > 0.5:
                image = (image[:, ::-1, :]).copy()
                depth_gt = (depth_gt[:, ::-1, :]).copy()

            # Random gamma, brightness, color augmentation
            do_augment = random.random()
            if do_augment > 0.5:
                image = self.augment_image(image)

        return image, depth_gt

    def augment_image(self, image):
        # gamma augmentation
        gamma = random.uniform(0.9, 1.1)
        image_aug = image ** gamma

        # brightness augmentation
        if self.config.dataset == 'nyu':
            brightness = random.uniform(0.75, 1.25)
        else:
            brightness = random.uniform(0.9, 1.1)
        image_aug = image_aug * brightness

        # color augmentation
        colors = np.random.uniform(0.9, 1.1, size=3)
        white = np.ones((image.shape[0], image.shape[1]))
        color_image = np.stack([white * colors[i] for i in range(3)], axis=2)
        image_aug *= color_image
        image_aug = np.clip(image_aug, 0, 1)

        return image_aug

    def __len__(self):
        return len(self.filenames)


class ToTensor(object):
    def __init__(self, mode, do_normalize=False, size=None):
        self.mode = mode
        self.normalize = transforms.Normalize(
            mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]) if do_normalize else nn.Identity()
        self.size = size
        if size is not None:
            self.resize = transforms.Resize(size=size)
        else:
            self.resize = nn.Identity()

    def __call__(self, sample):
        image, focal = sample['image'], sample['focal']
        image = self.to_tensor(image)
        image = self.normalize(image)
        image = self.resize(image)

        if self.mode == 'test':
            return {'image': image, 'focal': focal}

        depth = sample['depth']
        if self.mode == 'train':
            depth = self.to_tensor(depth)
            return {**sample, 'image': image, 'depth': depth, 'focal': focal}
        else:
            has_valid_depth = sample['has_valid_depth']
            image = self.resize(image)
            return {**sample, 'image': image, 'depth': depth, 'focal': focal, 'has_valid_depth': has_valid_depth,
                    'image_path': sample['image_path'], 'depth_path': sample['depth_path']}

    def to_tensor(self, pic):
        if not (_is_pil_image(pic) or _is_numpy_image(pic)):
            raise TypeError(
                'pic should be PIL Image or ndarray. Got {}'.format(type(pic)))

        if isinstance(pic, np.ndarray):
            img = torch.from_numpy(pic.transpose((2, 0, 1)))
            return img

        # 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))
        else:
            img = torch.ByteTensor(
                torch.ByteStorage.from_buffer(pic.tobytes()))
        # PIL image mode: 1, L, P, I, F, RGB, YCbCr, RGBA, CMYK
        if pic.mode == 'YCbCr':
            nchannel = 3
        elif pic.mode == 'I;16':
            nchannel = 1
        else:
            nchannel = len(pic.mode)
        img = img.view(pic.size[1], pic.size[0], nchannel)

        img = img.transpose(0, 1).transpose(0, 2).contiguous()
        if isinstance(img, torch.ByteTensor):
            return img.float()
        else:
            return img