Previously to zoom into an image I used the following (example) code:
import numpy as np
import cv2
image = np.zeros([64,64)]
image[10:-10,10:-10] = 1
# zoom into image:
imageZoomIn = cv2.resize(image[1:-1,1:-1] , (64,64))
#zoom out:
imageZoomOut = cv2.resize(image , (65,65))[1:-1, 1:-1)
#multiples Zooms:
imageBigZoom = image
for i in range(zoomScale):
imageBigZoom = cv2.resize(imageBigZoom [1:-1,1:-1] , (64,64))
Is ther a way to do this effectively in pytorch with whole batches? I want to train my NN with distorted data, which can be bigger or smaller than it is supposed to be.