Numpy: How to prevent circular looping to highest value when subtracting from the least value

Question

image = cv2.imread('black.jpg')[:,:,0] # let's just take the red channel alone for brevity
print(image[75:85,155:165]) # print middle 10x10
# output: [[0,0,0,....],...]

I want subtract 1 from all pixel intensity. And yes, I definitely do not want 0s to become 255s. I just want them to stay as 0.

print(image[75:85,155:165]-1)
# output: [[255,255,255,....],...]

print(np.array(image[75:85,155:165])-1)
# output: [[255,255,255,....],...]

print(np.array(image[75:85,155:165], dtype='float32')-1)
# output: [[-1.,-1.,-1.,....],...]

I can convert the last one back to uint8 after .clip(0,255) but that doesn't feel like the right way of doing it. Is there a way to directly do it (without casting and if conditions which might not be efficient for parallel processing)?

score 2 · Accepted Answer · answered Jun 06 '18 at 14:58

2

The best way to do it is np.where with mask:

grey_new = np.where(image == 0, 0, image - 1)

more info you can find in this answer, and numpy.where documentation.

answered Jun 06 '18 at 14:58

ykaner

1,740
1
9
13

this should be faster than subtracting and then pick things equal to 255 and making them 0, shouldn't it? – Saravanabalagi Ramachandran Jun 06 '18 at 15:25

score 0 · Answer 2 · answered Jun 06 '18 at 14:45

0

One easy option would be:

image[75:85,155:165] - 1 * (image[75:85,155:165] > 0)

answered Jun 06 '18 at 14:45

jdehesa

58,456
7
77
121

score 0 · Answer 3 · answered Jun 06 '18 at 14:46

0

Since you remove 1 to all values, only the 0 values will become 255. What you can do, then, is simply to put those back to 0 :

a = np.array(image[75:85,155:165])-1
a[a==255] = 0

answered Jun 06 '18 at 14:46

Ben

441
3
10

Numpy: How to prevent circular looping to highest value when subtracting from the least value

3 Answers3