Stitching

[Dave Allen]

Good morning,
I am trying to use Fiji to stitch together a large number of topographic images to create a wafer map of a silicon wafer.
I was wondering if anyone had any presentations or documents detailing how the stitching algorythms worked in Fiji

thanks
Dave

[Pavel Tomancak]

Hi Dave,

The documentation can be found on the Fiji wiki.

http://fiji.sc/wiki/index.php/Stitching_2D/3D

There is also a Bioinformatics paper, detailing the method.

S. Preibisch, S. Saalfeld, P. Tomancak (2009) "Globally optimal  
stitching of tiled 3D microscopic image acquisitions",Bioinformatics,  
25(11):1463-1465.

Its Open Access, i.e downloadable from here

http://bioinformatics.oxfordjournals.org/content/25/11/1463.abstract

All the best

PAvel

-----------------------------------------------------------------------------------
Pavel Tomancak, Ph.D.

Group Leader
Max Planck Institute of Molecular Cell Biology and Genetics
Pfotenhauerstr. 108
D-01307 Dresden Tel.: [phone]
Germany Fax: [phone]
[hidden email]
http://www.mpi-cbg.de
-----------------------------------------------------------------------------------




On Dec 9, 2011, at 12:27 PM, David Allen wrote:

> Good morning,
> I am trying to use Fiji to stitch together a large number of  
> topographic images to create a wafer map of a silicon wafer.
> I was wondering if anyone had any presentations or documents  
> detailing how the stitching algorythms worked in Fiji
>
> thanks
> Dave

[Vitali Khvatkov-2]

Dave, we have done quite a bit of stitching of wafer images from SEM
micros. (here is example: http://live.simagis.com/chipreverse-demo )

Your message does not say which modality you use, but if you use SEM
microscope there is number of specific issues that make this task more
challenging that stitching biological images (Digial Slides)
- there is no color information on SEM to assist stitching
- SEM drift may be considerable, so your image intensity is changing from
frame to frame
- Stage drift may be considerable, this is a problem in particular on
regular structures
- There may be overall geometry issues with acquisition where you image
array may not be rectangular.

There are some ImageJ / Fiji plugins that work with various degree of
success on biological samples. You can search this forum for "stitching" to
find number of post on this subject. In our tests none of those plugins
work well for SEM stitching of microchips.

If you do in-house solution some pointers are:
- don't rely too much on quality functions for alignment
- limit your alignment movements according to your stage precision
- watch for overall SEM array geometry

Best,

--
*Vitali Khvatkov*
Smart Imaging Technologies Co.
Office: [phone]
www.simagis.com


On Fri, Dec 9, 2011 at 5:27 AM, David Allen <[hidden email]> wrote:

> Good morning,
> I am trying to use Fiji to stitch together a large number of topographic
> images to create a wafer map of a silicon wafer.
> I was wondering if anyone had any presentations or documents detailing how
> the stitching algorythms worked in Fiji
>
> thanks
> Dave
>