Cropping Images

Cropping is a useful operation in computer vision that involves taking a subset of an image. Many applications allow user to crop an image on their profiles. Cropping is regularly used in building ML datasets as part of resampling and image to create variations of the same image for ML modelling purposes.

Cropping with Numpy

Index slicing is the easiest way to achieve cropping as demonstrated below

import cv2
import numpy as np
import matplotlib.pyplot as plt

%matplotlib inline
%config InlineBackend.figure_format = 'retina'

cat = cv2.imread('cat.jpg')
cat = cv2.cvtColor(cat, cv2.COLOR_BGR2RGB)

plt.imshow(cat)
plt.axis('off')
Example of Cropping an image with numpy 
cat.shape
(1280, 1920, 3)

Implementation with Numpy

To crop an image, we use numpy array slicing and pass the range of index location for the height and width to perform the crop. In the example below, the height range is $200:600$ and width range is 700:1090

cropped_cat = cat[ 200:600 , 700:1090 , : ]
plt.imshow(cropped_cat)
plt.axis('off')
Example of Cropping an image with numpy 
cropped_cat.shape
(400, 390, 3)