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
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
|
<!DOCTYPE html>
<!--[if IE 8]><html class="no-js lt-ie9" lang="en" > <![endif]-->
<!--[if gt IE 8]><!--> <html class="no-js" lang="en" > <!--<![endif]-->
<head>
<meta charset="utf-8">
<meta http-equiv="X-UA-Compatible" content="IE=edge">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<link rel="canonical" href="http://keras.io/preprocessing/image/">
<link rel="shortcut icon" href="../../img/favicon.ico">
<title>Image Preprocessing - Keras Documentation</title>
<link href='https://fonts.googleapis.com/css?family=Lato:400,700|Source+Sans+Pro:400,700|Inconsolata:400,700' rel='stylesheet' type='text/css'>
<link rel="stylesheet" href="../../css/theme.css" type="text/css" />
<link rel="stylesheet" href="../../css/theme_extra.css" type="text/css" />
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/9.12.0/styles/github.min.css">
<script>
// Current page data
var mkdocs_page_name = "Image Preprocessing";
var mkdocs_page_input_path = "preprocessing/image.md";
var mkdocs_page_url = "/preprocessing/image/";
</script>
<script src="../../js/jquery-2.1.1.min.js" defer></script>
<script src="../../js/modernizr-2.8.3.min.js" defer></script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/highlight.js/9.12.0/highlight.min.js"></script>
<script>hljs.initHighlightingOnLoad();</script>
<script>
(function(i,s,o,g,r,a,m){i['GoogleAnalyticsObject']=r;i[r]=i[r]||function(){
(i[r].q=i[r].q||[]).push(arguments)},i[r].l=1*new Date();a=s.createElement(o),
m=s.getElementsByTagName(o)[0];a.async=1;a.src=g;m.parentNode.insertBefore(a,m)
})(window,document,'script','https://www.google-analytics.com/analytics.js','ga');
ga('create', 'UA-61785484-1', 'keras.io');
ga('send', 'pageview');
</script>
</head>
<body class="wy-body-for-nav" role="document">
<div class="wy-grid-for-nav">
<nav data-toggle="wy-nav-shift" class="wy-nav-side stickynav">
<div class="wy-side-scroll">
<a href="">
<div class="keras-logo">
<img src="/img/keras-logo-small.jpg" class="keras-logo-img">
Keras Documentation
</div>
</a>
<div class="wy-side-nav-search">
<div role="search">
<form id ="rtd-search-form" class="wy-form" action="../../search.html" method="get">
<input type="text" name="q" placeholder="Search docs" title="Type search term here" />
</form>
</div>
</div>
<div class="wy-menu wy-menu-vertical" data-spy="affix" role="navigation" aria-label="main navigation">
<ul>
<li class="toctree-l1"><a class="reference internal" href="../..">Home</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../why-use-keras/">Why use Keras</a>
</li>
</ul>
<p class="caption"><span class="caption-text">Getting started</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../getting-started/sequential-model-guide/">Guide to the Sequential model</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../getting-started/functional-api-guide/">Guide to the Functional API</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../getting-started/faq/">FAQ</a>
</li>
</ul>
<p class="caption"><span class="caption-text">Models</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../models/about-keras-models/">About Keras models</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../models/sequential/">Sequential</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../models/model/">Model (functional API)</a>
</li>
</ul>
<p class="caption"><span class="caption-text">Layers</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../layers/about-keras-layers/">About Keras layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../layers/core/">Core Layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../layers/convolutional/">Convolutional Layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../layers/pooling/">Pooling Layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../layers/local/">Locally-connected Layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../layers/recurrent/">Recurrent Layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../layers/embeddings/">Embedding Layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../layers/merge/">Merge Layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../layers/advanced-activations/">Advanced Activations Layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../layers/normalization/">Normalization Layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../layers/noise/">Noise layers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../layers/wrappers/">Layer wrappers</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../layers/writing-your-own-keras-layers/">Writing your own Keras layers</a>
</li>
</ul>
<p class="caption"><span class="caption-text">Preprocessing</span></p>
<ul class="current">
<li class="toctree-l1"><a class="reference internal" href="../sequence/">Sequence Preprocessing</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../text/">Text Preprocessing</a>
</li>
<li class="toctree-l1 current"><a class="reference internal current" href="./">Image Preprocessing</a>
<ul class="current">
<li class="toctree-l2"><a class="reference internal" href="#imagedatagenerator-class">ImageDataGenerator class</a>
</li>
<li class="toctree-l2"><a class="reference internal" href="#imagedatagenerator-methods">ImageDataGenerator methods</a>
<ul>
<li class="toctree-l3"><a class="reference internal" href="#apply_transform">apply_transform</a>
</li>
<li class="toctree-l3"><a class="reference internal" href="#fit">fit</a>
</li>
<li class="toctree-l3"><a class="reference internal" href="#flow">flow</a>
</li>
<li class="toctree-l3"><a class="reference internal" href="#flow_from_dataframe">flow_from_dataframe</a>
</li>
<li class="toctree-l3"><a class="reference internal" href="#flow_from_directory">flow_from_directory</a>
</li>
<li class="toctree-l3"><a class="reference internal" href="#get_random_transform">get_random_transform</a>
</li>
<li class="toctree-l3"><a class="reference internal" href="#random_transform">random_transform</a>
</li>
<li class="toctree-l3"><a class="reference internal" href="#standardize">standardize</a>
</li>
</ul>
</li>
</ul>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../losses/">Losses</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../metrics/">Metrics</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../optimizers/">Optimizers</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../activations/">Activations</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../callbacks/">Callbacks</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../datasets/">Datasets</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../applications/">Applications</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../backend/">Backend</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../initializers/">Initializers</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../regularizers/">Regularizers</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../constraints/">Constraints</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../visualization/">Visualization</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../scikit-learn-api/">Scikit-learn API</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../utils/">Utils</a>
</li>
</ul>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../contributing/">Contributing</a>
</li>
</ul>
<p class="caption"><span class="caption-text">Examples</span></p>
<ul>
<li class="toctree-l1"><a class="reference internal" href="../../examples/addition_rnn/">Addition RNN</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/antirectifier/">Custom layer - antirectifier</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/babi_rnn/">Baby RNN</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/babi_memnn/">Baby MemNN</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/cifar10_cnn/">CIFAR-10 CNN</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/cifar10_resnet/">CIFAR-10 ResNet</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/conv_filter_visualization/">Convolution filter visualization</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/conv_lstm/">Convolutional LSTM</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/deep_dream/">Deep Dream</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/image_ocr/">Image OCR</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/imdb_bidirectional_lstm/">Bidirectional LSTM</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/imdb_cnn/">1D CNN for text classification</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/imdb_cnn_lstm/">Sentiment classification CNN-LSTM</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/imdb_fasttext/">Fasttext for text classification</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/imdb_lstm/">Sentiment classification LSTM</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/lstm_seq2seq/">Sequence to sequence - training</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/lstm_seq2seq_restore/">Sequence to sequence - prediction</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/lstm_stateful/">Stateful LSTM</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/lstm_text_generation/">LSTM for text generation</a>
</li>
<li class="toctree-l1"><a class="reference internal" href="../../examples/mnist_acgan/">Auxiliary Classifier GAN</a>
</li>
</ul>
</div>
</div>
</nav>
<section data-toggle="wy-nav-shift" class="wy-nav-content-wrap">
<nav class="wy-nav-top" role="navigation" aria-label="top navigation">
<i data-toggle="wy-nav-top" class="fa fa-bars"></i>
<a href="../..">Keras Documentation</a>
</nav>
<div class="wy-nav-content">
<div class="rst-content">
<div role="navigation" aria-label="breadcrumbs navigation">
<ul class="wy-breadcrumbs">
<li><a href="../..">Docs</a> »</li>
<li>Preprocessing »</li>
<li>Image Preprocessing</li>
<li class="wy-breadcrumbs-aside">
<a href="https://github.com/keras-team/keras/tree/master/docs"
class="icon icon-github"> Edit on GitHub</a>
</li>
</ul>
<hr/>
</div>
<div role="main">
<div class="section">
<h1 id="image-preprocessing">Image Preprocessing</h1>
<p><span style="float:right;"><a href="https://github.com/keras-team/keras/blob/master/keras/preprocessing/image.py#L238">[source]</a></span></p>
<h2 id="imagedatagenerator-class">ImageDataGenerator class</h2>
<pre><code class="python">keras.preprocessing.image.ImageDataGenerator(featurewise_center=False, samplewise_center=False, featurewise_std_normalization=False, samplewise_std_normalization=False, zca_whitening=False, zca_epsilon=1e-06, rotation_range=0, width_shift_range=0.0, height_shift_range=0.0, brightness_range=None, shear_range=0.0, zoom_range=0.0, channel_shift_range=0.0, fill_mode='nearest', cval=0.0, horizontal_flip=False, vertical_flip=False, rescale=None, preprocessing_function=None, data_format='channels_last', validation_split=0.0, interpolation_order=1, dtype='float32')
</code></pre>
<p>Generate batches of tensor image data with real-time data augmentation.
The data will be looped over (in batches).</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>featurewise_center</strong>: Boolean.
Set input mean to 0 over the dataset, feature-wise.</li>
<li><strong>samplewise_center</strong>: Boolean. Set each sample mean to 0.</li>
<li><strong>featurewise_std_normalization</strong>: Boolean.
Divide inputs by std of the dataset, feature-wise.</li>
<li><strong>samplewise_std_normalization</strong>: Boolean. Divide each input by its std.</li>
<li><strong>zca_epsilon</strong>: epsilon for ZCA whitening. Default is 1e-6.</li>
<li><strong>zca_whitening</strong>: Boolean. Apply ZCA whitening.</li>
<li><strong>rotation_range</strong>: Int. Degree range for random rotations.</li>
<li><strong>width_shift_range</strong>: Float, 1-D array-like or int<ul>
<li>float: fraction of total width, if < 1, or pixels if >= 1.</li>
<li>1-D array-like: random elements from the array.</li>
<li>int: integer number of pixels from interval
<code>(-width_shift_range, +width_shift_range)</code></li>
<li>With <code>width_shift_range=2</code> possible values
are integers <code>[-1, 0, +1]</code>,
same as with <code>width_shift_range=[-1, 0, +1]</code>,
while with <code>width_shift_range=1.0</code> possible values are floats
in the interval [-1.0, +1.0).</li>
</ul>
</li>
<li><strong>height_shift_range</strong>: Float, 1-D array-like or int<ul>
<li>float: fraction of total height, if < 1, or pixels if >= 1.</li>
<li>1-D array-like: random elements from the array.</li>
<li>int: integer number of pixels from interval
<code>(-height_shift_range, +height_shift_range)</code></li>
<li>With <code>height_shift_range=2</code> possible values
are integers <code>[-1, 0, +1]</code>,
same as with <code>height_shift_range=[-1, 0, +1]</code>,
while with <code>height_shift_range=1.0</code> possible values are floats
in the interval [-1.0, +1.0).</li>
</ul>
</li>
<li><strong>brightness_range</strong>: Tuple or list of two floats. Range for picking
a brightness shift value from.</li>
<li><strong>shear_range</strong>: Float. Shear Intensity
(Shear angle in counter-clockwise direction in degrees)</li>
<li><strong>zoom_range</strong>: Float or [lower, upper]. Range for random zoom.
If a float, <code>[lower, upper] = [1-zoom_range, 1+zoom_range]</code>.</li>
<li><strong>channel_shift_range</strong>: Float. Range for random channel shifts.</li>
<li><strong>fill_mode</strong>: One of {"constant", "nearest", "reflect" or "wrap"}.
Default is 'nearest'.
Points outside the boundaries of the input are filled
according to the given mode:<ul>
<li>'constant': kkkkkkkk|abcd|kkkkkkkk (cval=k)</li>
<li>'nearest': aaaaaaaa|abcd|dddddddd</li>
<li>'reflect': abcddcba|abcd|dcbaabcd</li>
<li>'wrap': abcdabcd|abcd|abcdabcd</li>
</ul>
</li>
<li><strong>cval</strong>: Float or Int.
Value used for points outside the boundaries
when <code>fill_mode = "constant"</code>.</li>
<li><strong>horizontal_flip</strong>: Boolean. Randomly flip inputs horizontally.</li>
<li><strong>vertical_flip</strong>: Boolean. Randomly flip inputs vertically.</li>
<li><strong>rescale</strong>: rescaling factor. Defaults to None.
If None or 0, no rescaling is applied,
otherwise we multiply the data by the value provided
(after applying all other transformations).</li>
<li><strong>preprocessing_function</strong>: function that will be applied on each input.
The function will run after the image is resized and augmented.
The function should take one argument:
one image (Numpy tensor with rank 3),
and should output a Numpy tensor with the same shape.</li>
<li><strong>data_format</strong>: Image data format,
either "channels_first" or "channels_last".
"channels_last" mode means that the images should have shape
<code>(samples, height, width, channels)</code>,
"channels_first" mode means that the images should have shape
<code>(samples, channels, height, width)</code>.
It defaults to the <code>image_data_format</code> value found in your
Keras config file at <code>~/.keras/keras.json</code>.
If you never set it, then it will be "channels_last".</li>
<li><strong>validation_split</strong>: Float. Fraction of images reserved for validation
(strictly between 0 and 1).</li>
<li><strong>dtype</strong>: Dtype to use for the generated arrays.</li>
</ul>
<p><strong>Examples</strong></p>
<p>Example of using <code>.flow(x, y)</code>:</p>
<pre><code class="python">(x_train, y_train), (x_test, y_test) = cifar10.load_data()
y_train = np_utils.to_categorical(y_train, num_classes)
y_test = np_utils.to_categorical(y_test, num_classes)
datagen = ImageDataGenerator(
featurewise_center=True,
featurewise_std_normalization=True,
rotation_range=20,
width_shift_range=0.2,
height_shift_range=0.2,
horizontal_flip=True)
# compute quantities required for featurewise normalization
# (std, mean, and principal components if ZCA whitening is applied)
datagen.fit(x_train)
# fits the model on batches with real-time data augmentation:
model.fit_generator(datagen.flow(x_train, y_train, batch_size=32),
steps_per_epoch=len(x_train) / 32, epochs=epochs)
# here's a more "manual" example
for e in range(epochs):
print('Epoch', e)
batches = 0
for x_batch, y_batch in datagen.flow(x_train, y_train, batch_size=32):
model.fit(x_batch, y_batch)
batches += 1
if batches >= len(x_train) / 32:
# we need to break the loop by hand because
# the generator loops indefinitely
break
</code></pre>
<p>Example of using <code>.flow_from_directory(directory)</code>:</p>
<pre><code class="python">train_datagen = ImageDataGenerator(
rescale=1./255,
shear_range=0.2,
zoom_range=0.2,
horizontal_flip=True)
test_datagen = ImageDataGenerator(rescale=1./255)
train_generator = train_datagen.flow_from_directory(
'data/train',
target_size=(150, 150),
batch_size=32,
class_mode='binary')
validation_generator = test_datagen.flow_from_directory(
'data/validation',
target_size=(150, 150),
batch_size=32,
class_mode='binary')
model.fit_generator(
train_generator,
steps_per_epoch=2000,
epochs=50,
validation_data=validation_generator,
validation_steps=800)
</code></pre>
<p>Example of transforming images and masks together.</p>
<pre><code class="python"># we create two instances with the same arguments
data_gen_args = dict(featurewise_center=True,
featurewise_std_normalization=True,
rotation_range=90,
width_shift_range=0.1,
height_shift_range=0.1,
zoom_range=0.2)
image_datagen = ImageDataGenerator(**data_gen_args)
mask_datagen = ImageDataGenerator(**data_gen_args)
# Provide the same seed and keyword arguments to the fit and flow methods
seed = 1
image_datagen.fit(images, augment=True, seed=seed)
mask_datagen.fit(masks, augment=True, seed=seed)
image_generator = image_datagen.flow_from_directory(
'data/images',
class_mode=None,
seed=seed)
mask_generator = mask_datagen.flow_from_directory(
'data/masks',
class_mode=None,
seed=seed)
# combine generators into one which yields image and masks
train_generator = zip(image_generator, mask_generator)
model.fit_generator(
train_generator,
steps_per_epoch=2000,
epochs=50)
</code></pre>
<p>Example of using <code>.flow_from_dataframe(dataframe, directory,</code>:</p>
<pre><code class="python">
train_df = pandas.read_csv("./train.csv")
valid_df = pandas.read_csv("./valid.csv")
train_datagen = ImageDataGenerator(
rescale=1./255,
shear_range=0.2,
zoom_range=0.2,
horizontal_flip=True)
test_datagen = ImageDataGenerator(rescale=1./255)
train_generator = train_datagen.flow_from_dataframe(
dataframe=train_df,
directory='data/train',
x_col="filename",
y_col="class",
target_size=(150, 150),
batch_size=32,
class_mode='binary')
validation_generator = test_datagen.flow_from_dataframe(
dataframe=valid_df,
directory='data/validation',
x_col="filename",
y_col="class",
target_size=(150, 150),
batch_size=32,
class_mode='binary')
model.fit_generator(
train_generator,
steps_per_epoch=2000,
epochs=50,
validation_data=validation_generator,
validation_steps=800)
</code></pre>
<hr />
<h2 id="imagedatagenerator-methods">ImageDataGenerator methods</h2>
<h3 id="apply_transform">apply_transform</h3>
<pre><code class="python">apply_transform(x, transform_parameters)
</code></pre>
<p>Applies a transformation to an image according to given parameters.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>x</strong>: 3D tensor, single image.</li>
<li><strong>transform_parameters</strong>: Dictionary with string - parameter pairs
describing the transformation.
Currently, the following parameters
from the dictionary are used:<ul>
<li><code>'theta'</code>: Float. Rotation angle in degrees.</li>
<li><code>'tx'</code>: Float. Shift in the x direction.</li>
<li><code>'ty'</code>: Float. Shift in the y direction.</li>
<li><code>'shear'</code>: Float. Shear angle in degrees.</li>
<li><code>'zx'</code>: Float. Zoom in the x direction.</li>
<li><code>'zy'</code>: Float. Zoom in the y direction.</li>
<li><code>'flip_horizontal'</code>: Boolean. Horizontal flip.</li>
<li><code>'flip_vertical'</code>: Boolean. Vertical flip.</li>
<li><code>'channel_shift_intencity'</code>: Float. Channel shift intensity.</li>
<li><code>'brightness'</code>: Float. Brightness shift intensity.</li>
</ul>
</li>
</ul>
<p><strong>Returns</strong></p>
<p>A transformed version of the input (same shape).</p>
<hr />
<h3 id="fit">fit</h3>
<pre><code class="python">fit(x, augment=False, rounds=1, seed=None)
</code></pre>
<p>Fits the data generator to some sample data.</p>
<p>This computes the internal data stats related to the
data-dependent transformations, based on an array of sample data.</p>
<p>Only required if <code>featurewise_center</code> or
<code>featurewise_std_normalization</code> or <code>zca_whitening</code> are set to True.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>x</strong>: Sample data. Should have rank 4.
In case of grayscale data,
the channels axis should have value 1, in case
of RGB data, it should have value 3, and in case
of RGBA data, it should have value 4.</li>
<li><strong>augment</strong>: Boolean (default: False).
Whether to fit on randomly augmented samples.</li>
<li><strong>rounds</strong>: Int (default: 1).
If using data augmentation (<code>augment=True</code>),
this is how many augmentation passes over the data to use.</li>
<li><strong>seed</strong>: Int (default: None). Random seed.</li>
</ul>
<hr />
<h3 id="flow">flow</h3>
<pre><code class="python">flow(x, y=None, batch_size=32, shuffle=True, sample_weight=None, seed=None, save_to_dir=None, save_prefix='', save_format='png', subset=None)
</code></pre>
<p>Takes data & label arrays, generates batches of augmented data.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>x</strong>: Input data. Numpy array of rank 4 or a tuple.
If tuple, the first element
should contain the images and the second element
another numpy array or a list of numpy arrays
that gets passed to the output
without any modifications.
Can be used to feed the model miscellaneous data
along with the images.
In case of grayscale data, the channels axis of the image array
should have value 1, in case
of RGB data, it should have value 3, and in case
of RGBA data, it should have value 4.</li>
<li><strong>y</strong>: Labels.</li>
<li><strong>batch_size</strong>: Int (default: 32).</li>
<li><strong>shuffle</strong>: Boolean (default: True).</li>
<li><strong>sample_weight</strong>: Sample weights.</li>
<li><strong>seed</strong>: Int (default: None).</li>
<li><strong>save_to_dir</strong>: None or str (default: None).
This allows you to optionally specify a directory
to which to save the augmented pictures being generated
(useful for visualizing what you are doing).</li>
<li><strong>save_prefix</strong>: Str (default: <code>''</code>).
Prefix to use for filenames of saved pictures
(only relevant if <code>save_to_dir</code> is set).</li>
<li><strong>save_format</strong>: one of "png", "jpeg"
(only relevant if <code>save_to_dir</code> is set). Default: "png".</li>
<li><strong>subset</strong>: Subset of data (<code>"training"</code> or <code>"validation"</code>) if
<code>validation_split</code> is set in <code>ImageDataGenerator</code>.</li>
</ul>
<p><strong>Returns</strong></p>
<p>An <code>Iterator</code> yielding tuples of <code>(x, y)</code>
where <code>x</code> is a numpy array of image data
(in the case of a single image input) or a list
of numpy arrays (in the case with
additional inputs) and <code>y</code> is a numpy array
of corresponding labels. If 'sample_weight' is not None,
the yielded tuples are of the form <code>(x, y, sample_weight)</code>.
If <code>y</code> is None, only the numpy array <code>x</code> is returned.</p>
<hr />
<h3 id="flow_from_dataframe">flow_from_dataframe</h3>
<pre><code class="python">flow_from_dataframe(dataframe, directory=None, x_col='filename', y_col='class', weight_col=None, target_size=(256, 256), color_mode='rgb', classes=None, class_mode='categorical', batch_size=32, shuffle=True, seed=None, save_to_dir=None, save_prefix='', save_format='png', subset=None, interpolation='nearest', validate_filenames=True)
</code></pre>
<p>Takes the dataframe and the path to a directory
and generates batches of augmented/normalized data.</p>
<p><strong>A simple tutorial can be found </strong><a href="http://bit.ly/keras_flow_from_dataframe">here</a>.</p>
<p><strong>Arguments</strong></p>
<ul>
<li>
<p><strong>dataframe</strong>: Pandas dataframe containing the filepaths relative to
<code>directory</code> (or absolute paths if <code>directory</code> is None) of the
images in a string column. It should include other column/s
depending on the <code>class_mode</code>:</p>
<ul>
<li>if <code>class_mode</code> is <code>"categorical"</code> (default value) it must
include the <code>y_col</code> column with the class/es of each image.
Values in column can be string/list/tuple if a single class
or list/tuple if multiple classes.</li>
<li>if <code>class_mode</code> is <code>"binary"</code> or <code>"sparse"</code> it must include
the given <code>y_col</code> column with class values as strings.</li>
<li>if <code>class_mode</code> is <code>"raw"</code> or <code>"multi_output"</code> it should contain</li>
</ul>
<p>the columns specified in <code>y_col</code>.</p>
<ul>
<li>if <code>class_mode</code> is <code>"input"</code> or <code>None</code> no extra column is needed.</li>
<li><strong>directory</strong>: string, path to the directory to read images from. If <code>None</code>,
data in <code>x_col</code> column should be absolute paths.</li>
<li><strong>x_col</strong>: string, column in <code>dataframe</code> that contains the filenames (or
absolute paths if <code>directory</code> is <code>None</code>).</li>
<li><strong>y_col</strong>: string or list, column/s in <code>dataframe</code> that has the target data.</li>
<li><strong>weight_col</strong>: string, column in <code>dataframe</code> that contains the sample
weights. Default: <code>None</code>.</li>
<li><strong>target_size</strong>: tuple of integers <code>(height, width)</code>, default: <code>(256, 256)</code>.
The dimensions to which all images found will be resized.</li>
<li><strong>color_mode</strong>: one of "grayscale", "rgb", "rgba". Default: "rgb".
Whether the images will be converted to have 1 or 3 color channels.</li>
<li><strong>classes</strong>: optional list of classes (e.g. <code>['dogs', 'cats']</code>).
Default: None. If not provided, the list of classes will be
automatically inferred from the <code>y_col</code>,
which will map to the label indices, will be alphanumeric).
The dictionary containing the mapping from class names to class
indices can be obtained via the attribute <code>class_indices</code>.</li>
<li><strong>class_mode</strong>: one of "binary", "categorical", "input", "multi_output",
"raw", sparse" or None. Default: "categorical".
Mode for yielding the targets:</li>
<li><code>"binary"</code>: 1D numpy array of binary labels,</li>
<li><code>"categorical"</code>: 2D numpy array of one-hot encoded labels.
Supports multi-label output.</li>
<li><code>"input"</code>: images identical to input images (mainly used to
work with autoencoders),</li>
<li><code>"multi_output"</code>: list with the values of the different columns,</li>
<li><code>"raw"</code>: numpy array of values in <code>y_col</code> column(s),</li>
<li><code>"sparse"</code>: 1D numpy array of integer labels,</li>
<li><code>None</code>, no targets are returned (the generator will only yield
batches of image data, which is useful to use in
<code>model.predict_generator()</code>).</li>
<li><strong>batch_size</strong>: size of the batches of data (default: 32).</li>
<li><strong>shuffle</strong>: whether to shuffle the data (default: True)</li>
<li><strong>seed</strong>: optional random seed for shuffling and transformations.</li>
<li><strong>save_to_dir</strong>: None or str (default: None).
This allows you to optionally specify a directory
to which to save the augmented pictures being generated
(useful for visualizing what you are doing).</li>
<li><strong>save_prefix</strong>: str. Prefix to use for filenames of saved pictures
(only relevant if <code>save_to_dir</code> is set).</li>
<li><strong>save_format</strong>: one of "png", "jpeg"
(only relevant if <code>save_to_dir</code> is set). Default: "png".</li>
<li><strong>follow_links</strong>: whether to follow symlinks inside class subdirectories
(default: False).</li>
<li><strong>subset</strong>: Subset of data (<code>"training"</code> or <code>"validation"</code>) if
<code>validation_split</code> is set in <code>ImageDataGenerator</code>.</li>
<li><strong>interpolation</strong>: Interpolation method used to resample the image if the
target size is different from that of the loaded image.
Supported methods are <code>"nearest"</code>, <code>"bilinear"</code>, and <code>"bicubic"</code>.
If PIL version 1.1.3 or newer is installed, <code>"lanczos"</code> is also
supported. If PIL version 3.4.0 or newer is installed, <code>"box"</code> and
<code>"hamming"</code> are also supported. By default, <code>"nearest"</code> is used.</li>
<li><strong>validate_filenames</strong>: Boolean, whether to validate image filenames in
<code>x_col</code>. If <code>True</code>, invalid images will be ignored. Disabling this
option can lead to speed-up in the execution of this function.
Default: <code>True</code>.</li>
</ul>
</li>
</ul>
<p><strong>Returns</strong></p>
<p>A <code>DataFrameIterator</code> yielding tuples of <code>(x, y)</code>
where <code>x</code> is a numpy array containing a batch
of images with shape <code>(batch_size, *target_size, channels)</code>
and <code>y</code> is a numpy array of corresponding labels.</p>
<hr />
<h3 id="flow_from_directory">flow_from_directory</h3>
<pre><code class="python">flow_from_directory(directory, target_size=(256, 256), color_mode='rgb', classes=None, class_mode='categorical', batch_size=32, shuffle=True, seed=None, save_to_dir=None, save_prefix='', save_format='png', follow_links=False, subset=None, interpolation='nearest')
</code></pre>
<p>Takes the path to a directory & generates batches of augmented data.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>directory</strong>: string, path to the target directory.
It should contain one subdirectory per class.
Any PNG, JPG, BMP, PPM or TIF images
inside each of the subdirectories directory tree
will be included in the generator.
See <a href="https://gist.github.com/fchollet/0830affa1f7f19fd47b06d4cf89ed44d">this script</a>
for more details.</li>
<li><strong>target_size</strong>: Tuple of integers <code>(height, width)</code>,
default: <code>(256, 256)</code>.
The dimensions to which all images found will be resized.</li>
<li><strong>color_mode</strong>: One of "grayscale", "rgb", "rgba". Default: "rgb".
Whether the images will be converted to
have 1, 3, or 4 channels.</li>
<li><strong>classes</strong>: Optional list of class subdirectories
(e.g. <code>['dogs', 'cats']</code>). Default: None.
If not provided, the list of classes will be automatically
inferred from the subdirectory names/structure
under <code>directory</code>, where each subdirectory will
be treated as a different class
(and the order of the classes, which will map to the label
indices, will be alphanumeric).
The dictionary containing the mapping from class names to class
indices can be obtained via the attribute <code>class_indices</code>.</li>
<li><strong>class_mode</strong>: One of "categorical", "binary", "sparse",
"input", or None. Default: "categorical".
Determines the type of label arrays that are returned:<ul>
<li>"categorical" will be 2D one-hot encoded labels,</li>
<li>"binary" will be 1D binary labels,
"sparse" will be 1D integer labels,</li>
<li>"input" will be images identical
to input images (mainly used to work with autoencoders).</li>
<li>If None, no labels are returned
(the generator will only yield batches of image data,
which is useful to use with <code>model.predict_generator()</code>).
Please note that in case of class_mode None,
the data still needs to reside in a subdirectory
of <code>directory</code> for it to work correctly.</li>
</ul>
</li>
<li><strong>batch_size</strong>: Size of the batches of data (default: 32).</li>
<li><strong>shuffle</strong>: Whether to shuffle the data (default: True)
If set to False, sorts the data in alphanumeric order.</li>
<li><strong>seed</strong>: Optional random seed for shuffling and transformations.</li>
<li><strong>save_to_dir</strong>: None or str (default: None).
This allows you to optionally specify
a directory to which to save
the augmented pictures being generated
(useful for visualizing what you are doing).</li>
<li><strong>save_prefix</strong>: Str. Prefix to use for filenames of saved pictures
(only relevant if <code>save_to_dir</code> is set).</li>
<li><strong>save_format</strong>: One of "png", "jpeg"
(only relevant if <code>save_to_dir</code> is set). Default: "png".</li>
<li><strong>follow_links</strong>: Whether to follow symlinks inside
class subdirectories (default: False).</li>
<li><strong>subset</strong>: Subset of data (<code>"training"</code> or <code>"validation"</code>) if
<code>validation_split</code> is set in <code>ImageDataGenerator</code>.</li>
<li><strong>interpolation</strong>: Interpolation method used to
resample the image if the
target size is different from that of the loaded image.
Supported methods are <code>"nearest"</code>, <code>"bilinear"</code>,
and <code>"bicubic"</code>.
If PIL version 1.1.3 or newer is installed, <code>"lanczos"</code> is also
supported. If PIL version 3.4.0 or newer is installed,
<code>"box"</code> and <code>"hamming"</code> are also supported.
By default, <code>"nearest"</code> is used.</li>
</ul>
<p><strong>Returns</strong></p>
<p>A <code>DirectoryIterator</code> yielding tuples of <code>(x, y)</code>
where <code>x</code> is a numpy array containing a batch
of images with shape <code>(batch_size, *target_size, channels)</code>
and <code>y</code> is a numpy array of corresponding labels.</p>
<hr />
<h3 id="get_random_transform">get_random_transform</h3>
<pre><code class="python">get_random_transform(img_shape, seed=None)
</code></pre>
<p>Generates random parameters for a transformation.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>seed</strong>: Random seed.</li>
<li><strong>img_shape</strong>: Tuple of integers.
Shape of the image that is transformed.</li>
</ul>
<p><strong>Returns</strong></p>
<p>A dictionary containing randomly chosen parameters describing the
transformation.</p>
<hr />
<h3 id="random_transform">random_transform</h3>
<pre><code class="python">random_transform(x, seed=None)
</code></pre>
<p>Applies a random transformation to an image.</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>x</strong>: 3D tensor, single image.</li>
<li><strong>seed</strong>: Random seed.</li>
</ul>
<p><strong>Returns</strong></p>
<p>A randomly transformed version of the input (same shape).</p>
<hr />
<h3 id="standardize">standardize</h3>
<pre><code class="python">standardize(x)
</code></pre>
<p>Applies the normalization configuration in-place to a batch of inputs.</p>
<p><code>x</code> is changed in-place since the function is mainly used internally
to standarize images and feed them to your network. If a copy of <code>x</code>
would be created instead it would have a significant performance cost.
If you want to apply this method without changing the input in-place
you can call the method creating a copy before:</p>
<p>standarize(np.copy(x))</p>
<p><strong>Arguments</strong></p>
<ul>
<li><strong>x</strong>: Batch of inputs to be normalized.</li>
</ul>
<p><strong>Returns</strong></p>
<p>The inputs, normalized.</p>
</div>
</div>
<footer>
<div class="rst-footer-buttons" role="navigation" aria-label="footer navigation">
<a href="../../losses/" class="btn btn-neutral float-right" title="Losses">Next <span class="icon icon-circle-arrow-right"></span></a>
<a href="../text/" class="btn btn-neutral" title="Text Preprocessing"><span class="icon icon-circle-arrow-left"></span> Previous</a>
</div>
<hr/>
<div role="contentinfo">
<!-- Copyright etc -->
</div>
Built with <a href="https://www.mkdocs.org/">MkDocs</a> using a <a href="https://github.com/snide/sphinx_rtd_theme">theme</a> provided by <a href="https://readthedocs.org">Read the Docs</a>.
</footer>
</div>
</div>
</section>
</div>
<div class="rst-versions" role="note" aria-label="versions">
<span class="rst-current-version" data-toggle="rst-current-version">
<a href="http://github.com/keras-team/keras/" class="fa fa-github" style="float: left; color: #fcfcfc"> GitHub</a>
<span><a href="../text/" style="color: #fcfcfc;">« Previous</a></span>
<span style="margin-left: 15px"><a href="../../losses/" style="color: #fcfcfc">Next »</a></span>
</span>
</div>
<script>var base_url = '../..';</script>
<script src="../../js/theme.js" defer></script>
<script src="../../search/main.js" defer></script>
<script type="text/javascript" defer>
window.onload = function () {
SphinxRtdTheme.Navigation.enable(true);
};
</script>
</body>
</html>
|
