Extract arbitrary rectangular patches from image in python

Question

Is there a way to extract a rectangle of my choice from an image, maybe using numpy arrays? Most implementations available seem to be for regular sliding window solutions, but those always include steps, or rectangles of the same aspect ratio, or something like that.

Is it possible to provide the beginning x and y coordinates and the width and height (or ending x and y coordinates), and extract exactly that rectangle? Can this be done using numpy arrays alone? Or is there another way to do this?

They were originally jpeg images and I've converted them into numpy arrays. — user961627
– user961627, Commented May 22, 2014 at 7:47
What exactly do you mean my extract, select 3 points manually? And what should your output look like? — jrsm
– jrsm, Commented May 22, 2014 at 7:47
By extract I mean for eg if I have a 360x240 image, and I want to extract a rectangle with starting point (10,10) and size (100,200), then the output should be a 100x200 sub-image which starts from x=10 and y=10. Not manually, I'll be providing the patch particulars in the code. — user961627
– user961627, Commented May 22, 2014 at 7:49
Do you mean something like rect = np.copy(image[10:110, 10:210]) — jrsm
– jrsm, Commented May 22, 2014 at 7:57

jrsm · Accepted Answer · 2014-05-22 14:05:33Z

3

The best way to do this would be slicing i.e.

rect = np.copy(img[y1:y1+height,x1:x1+width])

where (x1, y1) is the upper left corner of your rectangle.

answered May 22, 2014 at 14:05

jrsm

1,7152 gold badges23 silver badges41 bronze badges

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

Collectives™ on Stack Overflow

Extract arbitrary rectangular patches from image in python

1 Answer 1

Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related