A truth value error using numpy array

Question

I am working on changing the lightness of a grayscale image. However, I have problem using numpy array when I tried to add 100 to the each pixel value and set it to 255 when the value goes over 255

import numpy as np
import matplotlib.pyplot as plt
from scipy import misc

for i in range(0, img.shape[0]):
    for j in range(0, img.shape[1]):
        im2[i,j] = img[i,j] + 100
        if im2[i,j] > 255:
            im2[i,j] = 255     

plt.imshow(im2, cmap = 'gray')
plt.show()

these are the codes and I got an error message saying

if im2[i,j] > 255:
ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

Can anyone help me out with it?

Then that's your issue. You can't compare an array to a single value. — user3483203
– user3483203, Commented Mar 29, 2018 at 14:19
What are you trying to do with if im2[i, j] > 255: ? What's the purpose of this line ? — IMCoins
– IMCoins, Commented Mar 29, 2018 at 14:20
@chrisz you can compare arrays to a single value, but you get an array as a result — FlyingTeller
– FlyingTeller, Commented Mar 29, 2018 at 14:21

FlyingTeller · Accepted Answer · 2018-03-29 14:25:50Z

2

You can simply use the numpy clip function for this purpose:

im2 = np.clip(img + 100, 0, 255)

Explanation:
The first argument is the array you want to clip, in your case that is img with 100 added to each element. The second argument is the minnimum value, that should be 0. The third argument is the maximum value, in your case 255.

answered Mar 29, 2018 at 14:25

FlyingTeller

21k3 gold badges49 silver badges62 bronze badges

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

2 Comments

IMCoins Over a year ago

The version using np should be prefered to my solution imo.

Paul Panzer Over a year ago

This won't work if - as is often the case - the dtype is uint8. Better: np.clip(img, None, 155) + 100

Akhilesh · Accepted Answer · 2018-03-29 14:48:12Z

0

If you want, you can simplify the code is the following manner:

im2 = (np.asarray (img) + 100)
im2[im2>255] = 255

The first line will convert the img into numpy array (if img is list) and will add 255 in each element.

And the line will replace the value by 255 of im2 if it is greater than 255.

answered Mar 29, 2018 at 14:48

Akhilesh

1,0341 gold badge9 silver badges27 bronze badges

Comments

IMCoins · Accepted Answer · 2018-03-29 14:33:28Z

0

When you think about your error, The truth value of an array with more than one element is ambiguous, it seems logical.

You are trying to say Is my array (say [1, 2, 3]) superior to 255 ? Is its length above 255 ? Are you trying to know if the first element is superior ? The second ? The sum of the entire array? The root square of all the entire array ?

What you want to do, if I follow your comment correctly, is :

for idx in range(len(im2[i, j])):
    if im2[i, j][idx] > 255:
        im2[i, j][idx] = 255

edited Mar 29, 2018 at 14:33

answered Mar 29, 2018 at 14:28

IMCoins

3,3161 gold badge13 silver badges27 bronze badges

7 Comments

FlyingTeller Over a year ago

Your answer makes it sound like im2[i,j] > 255 is the same as len(im2[i,j]) > 255

IMCoins Over a year ago

@FlyingTeller , do you mean considering the code, or with the 2nd paragraph ?

finites Over a year ago

@IMCoins i want to compare the each single values in the array.

FlyingTeller Over a year ago

The start of the second paragraph makes it sound a bit like that. It might be helpful to point out that np.array([1,2,3,4]) > 2 will result in np.array([False, False, True, True]) and that one needs an additional for loop as you have suggested

FlyingTeller Over a year ago

@finites that is what the proposed for loop is for

|

Collectives™ on Stack Overflow

A truth value error using numpy array

3 Answers 3

2 Comments

Comments

7 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

3 Answers 3

2 Comments

Comments

7 Comments

Your Answer

Sign up or log in

Post as a guest

Related