Re:How L1-minimization can recover highly degraded images
by iskander on Tuesday February 23, 2010 @12:15AM
(#1144749)
>In practice, this doesn't work as well for more complicated images such as brains or abdomens. There is just much more to draw, and the images cease to be sparse.
Compressed sensing works just fine for brains, abdomens etc.. because the images need to be sparse in some basis (e.g. wavelets), not in the original space.
Re:How L1-minimization can recover highly degraded images
by yummyfajitas on Tuesday February 23, 2010 @12:28AM
(#1144770)
I haven't seen any good compressed sensing reconstruction for brains, though I'm not 100% up to date on the literature. Can you give me a cite?
The best I've seen is Candes/Donoho's reconstruction based on curvelets, but they don't offer any real improvement over regular reconstruction.
Re:How L1-minimization can recover highly degraded images
by iskander on Tuesday February 23, 2010 @02:10AM
(#1144926)
Check out Lustig's thesis. He finds that you can keep the same brain image quality while accelerating the sampling considerably.
http://yonit.stanford.edu/cgi-bin/axs/ax.pl?http://www-mrsrl...
Re:How L1-minimization can recover highly degraded images
by Schtarflucz on Tuesday February 23, 2010 @03:56AM
(#1145034)
Has someone been able to improve the (spatial or temporal) resolution of fMRI scans using this technique?
Re:How L1-minimization can recover highly degraded images
by yummyfajitas on Tuesday February 23, 2010 @09:03AM
(#1145363)
Thanks.
Imagine you have a pixel image, with missing data, and with noise. Imagine you know a-priori that the image was generated simply by placing a few white circles against a black background. Then data was lost/discarded, and white noise was added to the remaining data.
(The white circles are cross sections of the patient's blood vessels or some such feature (bile ducts, in this case), and contain a contrast agent. The contrast agent appears white on the MRI, everything else appears black. )
Now that you know the pixel image has such a simple form, you don't need to do fancy noise removal to clean it up. All you need to do is figure out where the white circles are. Once you know where the are, you can just redraw the image based on your calculated locations of circles instead of reprocessing the old one.
You can play the same trick in k-space, which gives you a potent MRI reconstruction algorithm.
In practice, this doesn't work as well for more complicated images such as brains or abdomens. There is just much more to draw, and the images cease to be sparse.