Python Deep Learning Projects

Generating low-resolution images

To generate low-resolution images, define a function called reshape() that will resize the input image/digit to size 14*14. After defining this, we will use the reshape() function to generate low-resolution train and test images:

def reshape(x):    """Reshape images to 14*14"""    img = imresize(x.reshape(28,28), (14, 14))    return img# create 14*14 low resolution train and test imagesXX_train = np.array([*map(reshape, X_train.astype(float))])XX_test = np.array([*map(reshape, X_test.astype(float))])

XX_train and XX_test will be the images that we will feed into the encoder, and X_train and X_test will be the targets.