Welcome Guest! To enable all features, please Login or Register.

Notification

Icon
Error

Options
View
Last Go to last post Unread Go to first unread post
#1 Posted : Sunday, November 9, 2008 8:44:01 PM(UTC)
programsunlimited

Groups: Registered
Posts: 6


I have Leadtools Medical Imaging v 15 and am working on a project where one of the requirements is to automatically stitch together a mosaic of overlapping images and present a single combined image for viewing.

There are commercial packages like PanaVue and AutoStitch which automate this process... We'd like to use Leadtools to do this with overlapping shots of x-ray images.

I have tried the simple image stitching examples but am looking for something that will automatically warp, resample and blend together the images. I believe that I have to use some method to detect key points in common between the overlapping images and then measure distances and combine.
 

Try the latest version of LEADTOOLS for free for 60 days by downloading the evaluation: https://www.leadtools.com/downloads

Wanna join the discussion? Login to your LEADTOOLS Support accountor Register a new forum account.

#2 Posted : Monday, November 10, 2008 4:19:03 AM(UTC)

Qasem Lubani  
Guest

Groups: Guests
Posts: 3,028

Was thanked: 2 time(s) in 2 post(s)

Is it possible to send me a sample image that shows the original files and how the result after stitching the image should be? Please put the images in a ZIP or RAR file prior to sending them.

You can either post the files here, or send them to support@leadtools.com
 
#3 Posted : Monday, November 10, 2008 8:06:23 AM(UTC)
programsunlimited

Groups: Registered
Posts: 6


Here is a sample of the images. The images are labelled Row1Col.. and Row2Col.. where the Col in the name refers to the position of the image in the row.

Looking at the attachment the images correspond to these positions

Row1Col1 Row1Col2 Row1Col3 Row1Col4 Row1Col5
Row2Col1 Row2Col2 Row2Col3 Row2Col4 Row2Col5

Each image has some overlap to the image next to it and the image below it.

I'd like to be able to combine the images in a single image that looks like the FullImage.jpg in the attachment.
 
#4 Posted : Monday, November 10, 2008 8:09:14 AM(UTC)
programsunlimited

Groups: Registered
Posts: 6


For some reason the zip file did not attach to the post. I'm trying to attach it here and will send to support@leadtools.com as well.
File Attachment(s):
Xray.zip (112kb) downloaded 35 time(s).
 
#5 Posted : Tuesday, November 11, 2008 7:05:51 AM(UTC)

Qasem Lubani  
Guest

Groups: Guests
Posts: 3,028

Was thanked: 2 time(s) in 2 post(s)

A useful function that might help you improve the results when the bordering images do not have the exact same angle and scale is the Combine Warp function.

However, even with this function, making this process 100% automatic using LEADTOOLS will not be an easy thing to do. Your main problem will be defining at least 2 identical points in each of the 2 adjacent tiles. You may have to do this with help from the user if the amount of shift is not fixed every time.
Once you have these points, you can begin with an image as the foundation and combine it with a big empty background image without warping, then start to add the other tiles with the warp function so that they fit in with the previously added tiles.
 
#6 Posted : Tuesday, November 11, 2008 10:40:15 AM(UTC)
programsunlimited

Groups: Registered
Posts: 6


Thanks for your response. I was wondering if the L_CorrelationListBitmap and L_CorrelationBitmap together with L_CombineBitmapWarp and/or L_CombineBitmap could give a reasonable level of automatic point detection? If so, are these available in .NET code or only through C DLLs.
 
#7 Posted : Thursday, November 13, 2008 7:03:27 AM(UTC)

Adnan Ismail  
Guest

Groups: Guests
Posts: 3,028

Was thanked: 2 time(s) in 2 post(s)

Using the correlation function, you might be able to locate similar points but you must extract small rectangular areas from one image and compare them to the other image. Also, you should test this approach thoroughly with different cases, because the function is not intelligent enough to work with variations in images. This means you need to specify threshold values suitable for your images that are neither too high nor too low to get correct results.
The name of this function in .NET is CorrelationCommand Class.
 
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.

Powered by YAF.NET | YAF.NET © 2003-2020, Yet Another Forum.NET
This page was generated in 0.167 seconds.