Image resize do not display the correct image in opencv python

Question

I am using Opencv 4.1.0 with python 3.

I resized a 480 X 640 size original image to 500 x 500 that works fine.

Again resized the 500 X 500 image to 25000 X 1 . I am doing this for the purpose of facial recognition using PCA.

Again when I resized the image from 25000 X 1 to 500 X 500 , it did not display the 500 X 500 image . It instead displayed a error output of black and white vertical lines.

Can someone please point out what is going wrong here ? Thanks in advance.

import cv2
import numpy as np
image=cv2.imread('C://Users//raghu//Documents//Faces//gt_db//unique//IMG1.jpg',0)
I1=cv2.resize(image,(500,500))
I2=cv2.resize(I1,(25000,1))
I3=cv2.resize(I2,(500,500))
cv2.imshow('480x640',image)
cv2.imshow('25000X1',I2)
cv2.imshow('500x500',I1)
cv2.imshow('2-500X500',I3)
cv2.waitKey(0)
cv2.destroyAllWindows()

Resizing to I1 works fine.Expecting the output of I3 same as I1.

That is to be expected. When you resize to 1 row or column, you lose information, which cannot be recovered by resizing larger again. — fmw42
– fmw42, Commented Aug 21, 2019 at 23:16

J.D. · Accepted Answer · 2019-08-21 23:03:51Z

2

By downsizing to 1 pixel you loose nearly all image information as all y-pixels get interpolated to a single number per x-pixel. By resizing back that pixel is then copied vertically to 500px, so I expect you to get a stripy pattern.

You should not resize, you have to reshape. That means putting the pixel values from a 2d array to a 1d array, that is what the PCA algorithm expects.

Example:

# create 2d array
y = np.array(range(9)).reshape(3,3)
print(y)

[[0 1 2]
[3 4 5]
[6 7 8]]

# reshape to 1d
x = y.reshape(-1)
print(x)
print(x.shape[:2])

[0 1 2 3 4 5 6 7 8]
(9,)

reshape documentation

edited Aug 21, 2019 at 23:03

answered Aug 21, 2019 at 22:56

J.D.

4,5442 gold badges10 silver badges23 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Image resize do not display the correct image in opencv python

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related