Help regarding optimization tool box for image processing?

Hello I have two images a and b, where b is a rotated and translated image of a. Using Sum of Squared Difference method, I have to find the minimization function using optimization tool box by taking 8 by 8 box and running iterations. How do I go about it?

댓글 수: 5

Write down the function you want to minimize in equation form. Then we can offer advice about how to code it.
Emmanuel
Emmanuel 2014년 10월 11일
편집: Emmanuel 2014년 10월 11일
Hi! Yeah sure ! Thank you!
cost function = Summation [f(x) - g'(x))^2]
g' is the warped image of f and x is the pixel point. I have to select 8x8 block of pixels, optimize the cost function for n number of iterations.
Can you guide me as to how to code this in optimization toolbox?
So each 8x8 block must have its own distinct warp parameters? What it the form of the warping? Elastic, affine, rigid?
Please post your two images so we don't have to guess, and can give advice faster .
The warping consits of translation and rotation .
The original picture is :
The warped picture is:

댓글을 달려면 로그인하십시오.

 채택된 답변

Image Analyst
Image Analyst 2014년 10월 9일

2 개 추천

Why not just us imregister() which is in the Image Processing Toolbox?

댓글 수: 5

Yeah! But, I must implement it using optimization tool box. Can you please help me with that?
No. I'm sorry but I don't have that toolbox.
See http://www.mathworks.com/help/vision/examples/find-image-rotation-and-scale-using-automated-feature-matching.html, though your example is pretty substantially different so I don't know how well it would work.
Sure, you can frame it as an optimization problem where you minimize the error of the registered images but that's like using a crowbar to tear down a house. It's possible, but it's going to be a lot more work than using a bull dozer.
@Image Analyst Thanks a lot !!!!!!! It works!
@Sean de Wolski Yeah you are right! I am new to Image proc and yet to learn stuffs! Thanks a lot

댓글을 달려면 로그인하십시오.

추가 답변 (0개)

카테고리

도움말 센터File Exchange에서 Computer Vision Toolbox에 대해 자세히 알아보기

질문:

2014년 10월 9일

댓글:

2014년 10월 15일

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by