Thread Subject: Register two images to one another

I have a question about simultaneously registering two images to one another.

Suppose I have two images imA and imB that overlap somewhat (e.g. they’re two halves of a panorama). I want to register them. If I follow the Matlab Image Registration example, I get one that stays unmodified (the base image), and one that gets registered to it, often getting distorted a lot in the case of a projective or affine transformation.

So, is there a way to calculate two transforms AB and BA to register A to B and B to A respectively, then simply apply part of those two transformations so that the images “meet in the middle” so to speak? Do I just divide the two transformation matrices in two?

Thanks.

"Simon " <spsilk@gmail.com> wrote in message <hnm6g1$49$1@fred.mathworks.com>...
> I have a question about simultaneously registering two images to one another.
>
> Suppose I have two images imA and imB that overlap somewhat (e.g. they’re two halves of a panorama). I want to register them. If I follow the Matlab Image Registration example, I get one that stays unmodified (the base image), and one that gets registered to it, often getting distorted a lot in the case of a projective or affine transformation.
>
> So, is there a way to calculate two transforms AB and BA to register A to B and B to A respectively, then simply apply part of those two transformations so that the images “meet in the middle” so to speak? Do I just divide the two transformation matrices in two?
================

Since the reasons for the distortion are unclear (you've said little about them and why you think they occur), so are the benefits of finding an average transformation.

That said, there is some image registration literature in which, for similarity measure sim(A,B) and transform operator T, people try to minimize

sim(T*A,B)+sim(A,inv(T)*B)

I don't know of any such registration algorithm offered by the Image Processing Toobox, though, so you'd probably have to code it yourself.

Simon:
Here you go:
http://www.bing.com/search?q=image+stitching+algorithm&go=&form=QBLH&qs=n
The first link from Microsoft is a good one:

Image Alignment and Stitching: A Tutorial

Richard Szeliski
October 2004

This tutorial reviews image alignment and image stitching algorithms.
Image alignment (registration) algorithms can discover the large-scale
(parametric) correspondence relationships among images with varying
degrees of overlap. They are ideally suited for applications such as
video stabilization, summarization, and the creation of large-scale
panoramic photographs. Image stitching algorithms take the alignment
estimates produced by such registration algorithms and blend the
images in a seamless manner, taking care to deal with potential
problems such as blurring or ghosting caused by parallax and scene
movement as well as varying image exposures. This tutorial reviews the
basic motion models underlying alignment and stitching algorithms,
describes effective direct (pixel-based) and feature-based alignment
algorithms, and describes blending algorithms used to produce seamless
mosaics. It closes with a discussion of open research problems in the
area.

Here's a free registration package:

Registax V5 software for alignment/stacking/processing of images
http://www.astronomie.be/registax/