Hacker News new | past | comments | ask | show | jobs | submit login

>The trick with overlaying follow-up scans is called image fusion and is easy and can be done by one-click applications

Just wanted to point out that image fusion is far from easy. The brain is easier than most areas of the body because the skull provides a good basis for performing a registration, but even then the brain has some small room to shift and you can't get a perfect alignment with only rigid transformations, especially when you're dealing with tumors.

For most other parts of the body you have complications like organs shifting around, gas passing through the digestive system, patients being in slightly different orientations, etc. In these cases you most definitely need to use some kind of deformable registration, which is far from a solved problem, especially when you have things like tumors changing shape/size or even appearing from nowhere from one scan to the next.




Yes it was a bit of a generalization, but i mean the practical use of the tool. The technology behind it is wonderful! For my specialty its relatively easy (bone being a rigid body). But this type of software already works quite well for lung, liver etc. With careful manual alignment it can work for soft tissue sarcoma (muscle) and other types of more movable tissue, providing a quick overview. Volume measurement is also an option. But as you say its always complimentary to standard workflow.


Yeah, I wasn't trying to be disagreeable, just pointing out that this is a very active area of research still, and there is always room for improvement. I work mostly with segmentation and shape analysis instead of registration, but I am involved with a project on multi-modal image fusion, which is even harder (and more interesting).


I've done a lot of work on multi-modal registration and would be interested in talking to you about this. It's an amazingly hard problem and I see a lot of wheels being reinvented whenever I look at what other people are doing.

My own work has been mostly commercial and so under NDA or otherwise unpublishable, but is primarily based on applications of the pseudo-correlation algorithm (Radcliffe, Rajapakshe, Shalev, Medical Physics, vol. 21, No. 6, pp. 761 769, June 1994.) I can be reached at tradcliffe at predictivepatterns.com if you're interested.


Awesome, we use a lot of PET/CT/MRI fusion imaging for computer assisted surgery and have an EOS scanner (statistical shape modeling). Just really awesome tech. Good ultrasound -> CT fusion would, I think, be indeed be a major breakthrough, allowing easy matching / guided procedures everywhere in the body but also for us as a secondary check on CAS / robotic accuracy.


Even without physiological changes, multi-modal fusion is a made more difficult by non-rigid spatial transforms inherent to some of the modes (e.g. CT quite accurate spatially, but MR is certainly not, and can vary with pulse sequence also)




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: