Cannot import name 'load_img' from 'keras.preprocessing.image'

Question

Why am I getting this issue? I can import image module from kera.preprocessing. But cannot import image_dataset_from_directory. My TensorFlow version is 2.9.1, so I am not dealing with an old version, e.g. ImportError: cannot import name 'image_dataset_from_directory' from 'tensorflow.keras.preprocessing' (unknown location)

# make a prediction for a new image.
from keras.preprocessing.image import load_img
from keras.preprocessing.image import img_to_array
from keras.models import load_model

# load and prepare the image
def load_image(filename):
    # load the image
    img = load_img(filename, target_size=(224, 224))
    # convert to array
    img = img_to_array(img)
    # reshape into a single sample with 3 channels
    img = img.reshape(1, 224, 224, 3)
    # center pixel data
    img = img.astype('float32')
    img = img - [123.68, 116.779, 103.939]
    return img

# load an image and predict the class
def run_example():
    # load the image
    img = load_image('test.jpg')
    # load model
    model = load_model('final_model.h5')
    # predict the class
    result = model.predict(img)
    print(result[0])

# entry point, run the example
run_example()

error

from keras.preprocessing.image import load_img
ImportError: cannot import name 'load_img' from 'keras.preprocessing.image'

Answer 1

keras.preprocessing API is deprecated in Tensorflow 2.9.1. Please use tf.keras.utils instead, to import load_img as follows:

from tensorflow.keras.utils import load_img

To load dataset from directories please use tensorflow.keras.utils.image_dataset_from_directory. For more details, please refer to this link.