简体   繁体   中英

Getting ValueError: setting an array element with a sequence from tf.contrib.keras.preprocessing.image.ImageDatagenerator.flow

I am trying to do Data Augmentation in Tensorflow. I have written this code.

import numpy as np
import tensorflow as tf
import tensorflow.contrib.keras as keras
import time, random

def get_image_data_generator():
    return keras.preprocessing.image.ImageDataGenerator(
    rotation_range=get_random_rotation_angle(),\
    width_shift_range=get_random_wh_shift(),\
    height_shift_range=get_random_wh_shift(),\
    shear_range=get_random_shear(),\
    zoom_range=get_random_zoom(),\
    horizontal_flip=get_random_flip(),\
    vertical_flip=get_random_flip(),\
    preprocessing_function=get_random_function())

def augment_data(image_array,label_array):
    print image_array.shape
    images_array = image_array.copy()
    labels_array = label_array.copy()
    #Create a list of various datagenerators with different arguments
    datagenerators = []
    ndg = 10
    #Creating 10 different generators
    for ndata in xrange(ndg):
        datagenerators.append(get_image_data_generator())
    #Setting batch_size to be equal to no.of images
    bsize = image_array.shape[0]
    print bsize
    #Obtaining the augmented data
    for dgen in datagenerators:
        dgen.fit(image_array)
        (aug_img,aug_label) = dgen.flow(image_array,label_array,batch_size=bsize,shuffle=True)
        print aug_img.shape
        #Concatenating with the original data
        images_array = np.concatenate([images_array,aug_img],axis=0)
        labels_array = np.concatenate([labels_array,aug_label],axis=0)
    return (images_array,labels_array)

When I run the code using

augment_data(image_array,label_array)

I get an error which says

Traceback (most recent call last):
  File "cnn_model.py", line 40, in <module>
    images_array,labels_array = augment_data(image_array,label_array)
  File "/media/siladittya/d801fb13-809a-41b4-8f2e-d617ba103aba/ISI/code/2. known_object_detection/aug_data.py", line 47, in augment_data
    (aug_img,aug_label) = dgen.flow(image_array,label_array,batch_size=10000,shuffle=True)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/keras/_impl/keras/preprocessing/image.py", line 1018, in next
    return self._get_batches_of_transformed_samples(index_array)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/keras/_impl/keras/preprocessing/image.py", line 991, in _get_batches_of_transformed_samples
    batch_x[i] = x
ValueError: setting an array element with a sequence.

Edit :: I am getting this error even if I pass a single image as argument.

What am I doing wrong here? I can't understand. Please help.

Edit :: I am getting this error even if I pass a single image as argument.`

Can you pass the single element as an array and see:

example:

image_array, label_array = augment_data([image], [label])

The technical post webpages of this site follow the CC BY-SA 4.0 protocol. If you need to reprint, please indicate the site URL or the original address.Any question please contact:yoyou2525@163.com.

 
粤ICP备18138465号  © 2020-2024 STACKOOM.COM