MNIST saving images from CSV using matplotlib

Question

In the MNIST dataset, I have the images in the CSV format, each of the 784 columns corresponds to a pixel intensity. I want to save each of these images without looking at them with imshow.

import numpy as np
import csv
import matplotlib.pyplot as plt
i=0
with open('Book1.csv', 'r') as csv_file:
for data in csv.reader(csv_file):

    # The rest of columns are pixels
    pixels = data[:]

    # This array will be of 1D with length 784
    # The pixel intensity values are integers from 0 to 255
    pixels = np.array(pixels, dtype='uint8')

    # Reshape the array into 28 x 28 array (2-dimensional array)
    pixels = pixels.reshape((28, 28))
    i +=1
    # Plot
    plt.title('Label is {label}'.format(label=label))
    plt.imshow(pixels, cmap='gray')
    plt.savefig(str(i))'

With this I am unable to save each image.

Answer 1

The name of the file should end with the extention you want,

plt.savefig(str(i)+".png")

or you need to specify the format:

plt.savefig(str(i), format="png")

One additional problem you may run into is that after a while your plotting slows down, because all images are placed on top of each other. To circumvent this, you may call plt.close() at the end of the loop.