Ask Your Question
2

how to extract similar part of two images

asked 2014-05-28 05:32:04 -0600

nqthiep gravatar image

updated 2014-05-30 02:20:26 -0600

Hi all,

I have two images as below:

image description image description

I want to ask, how to extract similar part of there images.

Thanks for helping me.

=========================================================================================== Thank you very much for reply, Witek & GilLevi.

I have solved this problem in my way.

Here is my solution: - Get the end part of the first image, find its position in second image. Temporary called result A. - Get the first part of the second image, find its position in first image. Temporary called result B.

=> The region between A and B is result which i want.

edit retag flag offensive close merge delete

Comments

1

If you assume horizontal motion of the camera, that will work. In general case of finding similar parts present in both images - definitely not.

Witek gravatar imageWitek ( 2014-05-30 03:48:07 -0600 )edit

Thanks Witek. Yes, it is. I work on horizontal motion.

nqthiep gravatar imagenqthiep ( 2014-06-05 21:53:41 -0600 )edit

2 answers

Sort by ยป oldest newest most voted
1

answered 2014-05-29 03:46:04 -0600

Witek gravatar image

A friend of mine did some research on it. Unfortunately I am not sure if I can redistribute his papers. Look here for some interesting publications: http://www.researchgate.net/profile/Andrzej_Sluzek/publications

edit flag offensive delete link more
1

answered 2014-05-28 20:22:09 -0600

One method of doing so would be to detect keypoints in the two images, extract descriptors and match them. Then, you can take the bounding box of the matched keypoints in the two images.

edit flag offensive delete link more

Comments

1

You can use SURF or SIFT ... to detect keypoints and then match them. There are many good implementations available in OpenCV for SIFT , SURF...

r_asmak gravatar imager_asmak ( 2014-05-31 12:17:08 -0600 )edit

Question Tools

1 follower

Stats

Asked: 2014-05-28 05:32:04 -0600

Seen: 400 times

Last updated: May 30 '14