-
Notifications
You must be signed in to change notification settings - Fork 0
/
Image_Preprocess.py
39 lines (28 loc) · 1.16 KB
/
Image_Preprocess.py
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
''' Resizing the images and Storing them as numpy array '''
import cv2
import numpy as np
from numpy import asarray
from numpy import save
from tensorflow.keras.preprocessing.image import load_img
from tensorflow.keras.preprocessing.image import img_to_array
import os
os.chdir('D:/.../Project/Images') # Home directory of images
n=int(input("Enter num of images for Training and Val:"))
folder = 'D:/.../Project/Resized_Images' # New directory for storing resized images
for img in range(1,n+1):
img_data=cv2.imread(str(img)+'.png')
new_img=cv2.resize(img_data, (32,64), interpolation = cv2.INTER_AREA) # Resizing in 64x32 format
cv2.imwrite(os.path.join(folder , str(img)+'.png'), new_img)
print("....Resizing Complete....")
os.chdir(folder)
# Storing in a numpy format
photos= list()
for file in range(1,n+1):
print(file)
photo = cv2.imread(str(file)+'.png')
# photo = load_img(str(file)+'.png') # Can Add: target_size=(200, 200)
photo = img_to_array(photo) # Convert to numpy array
photos.append(photo)
photos = asarray(photos) # Storing into a singular array file
print(photos.shape)
save('Digits_Photo.npy', photos)