PDA

View Full Version : Advises on thesis development on Medical/Bylogical Image Processing with Qt.



danielperaza
31st October 2010, 18:35
Hi there Qt fans!.

I'm looking for a topic for my computer science's degree thesis. One of them, is about building a Tool for Biological/Medical Image processing (think of a kind of photoshop for processing images acquired through electronic microscopes). I really like this topic.

However, I think I'll have to use the Graphics View Framework which I don't have any experience with. I fear this forces me to invest a lot of time reading some docs instead of making some progress with my code.

On the other hand, is the fact that Qt doesn't provide too much code to work on image processing. So I think I'll have to use Qt along with OpenCV, CImg, ImageMagick or some other framework like these, and converting data structures back and forth between them in order to build my GUI with Qt.

Please, I need some advise on these two problems, since I really need to get my thesis done very very quickly (say 6 months or earlier working part-time), and I don't want my graduation date to be postponed. If this topic gets too much complicated, I'll start searching for another one.

SixDegrees
31st October 2010, 18:48
You're correct that Qt doesn't off much in the way of image processing software. It's meant primarily as a GUI library.

That said, the QGraphicsView/Scene/Item framework is excellent for displaying images. Learning how to use it may take you a couple of days, but that time will be repaid when time comes to do things like zooming, centering, translations, rotations, scaling and so on. You can build a very nice frontend for an image processing program in a matter of hours with Qt instead of writing all manner of tedious code yourself.

ImageMagick and OpenCV are both good choices for actual image manipulation and processing. Again, both take some time to get up to speed on, but that time will be repaid in overall savings in coding time. I have never used ImageMagick's C API, only the command-line interface, but it offers a lot. We use OpenCV at work quite a bit, and I'd much rather spend an hour or two learning how, say, to use an existing Kalman filtering routine that someone else has already written and debugged, than spend a week or two writing my own and getting much of it wrong.

As far as I know, there isn't any single package that handles both problems well; they are two essentially separate fields of study. So you'll either have to write a signficant amount of code to do what you want, or you'll wind up using two (possibly more) separate external packages like Qt and OpenCV anyway.

srsr
31st October 2010, 19:18
May be you can check at these links:

http://ftp.elvitec.fr/Elvitec/Dlls-Utilitaires-Elvitec/Driver_DSDX/Utilitaires%20DirectX-DirectShow/
http://ftp.elvitec.fr/Pleora/TECHNOTES-DEMOS/PureGEV_QtSample/

I'ts a Qt application with camera acquisition and OpenCV / IPP image processing.
link1 is for Qt acquiring from DirectShow camera
link2 is for Qt acquiring from GenICam cameras
Both use same OpenCV Plugin layer that you can already (re)use.
Hope this can help you in regards to your project : stay with Qt.
SRSR