Chris' Random Thoughts
|Jan. 10th, 2011 11:41 am MRI on the brain|
Once again, my mind is too active at night.
Current MRI scanners are much less accurate than they could be, and slower than they should be. The current models use relatively few antennas (4 to 32), and that leaves too much phase confusion in the reconstructed image -- aka blurring and noise. I think that MRI needs to be approached more like a phased array or synthetic aperture receiver system: many antennas (hundreds or more) arranged in a grid pattern to resolve the large number of point sources and complex phase relationships. With a greater number of antennas and better sampling, we should be able to gather multiple frequency slices at one time -- covering a larger volume in a single RF pulse. And this improved precision should help even more with pulsed gradient and relaxation time measurements.
What I'm not sure about is the best antenna grid arrangement. The obvious rectangular grid near the imaging region would most likely distort the magnetic field, and cause a lot of mechanical stress on the antenna array. The same grid might work if placed outside the imaging region, but that would increase distance, reduce signal strength, reduce position accuracy, and allow more noise sources to affect the signal. A cylindrical grid near the imaging region seems like it would distort less, reduce noise and signal strength issues, and not have as much mechanical stress -- but I'm not sure about the solution of the image. The solution for a rectangular grid is an easy FFT. But a cylindrical grid would need a mixed basis solution (part FFT, part something else).
Hmm, existing MRI simulator packages look like they won't help here. I may have to write something in my "copious free time".
Current Mood: ThinkingLeave a comment
|Oct. 21st, 2010 10:07 am Sometimes I just can't go back to sleep|
This is a physics thing: if you don't recognize the equations, you can safely skip this post.
F = ma
That's fine in S.I. units, but in English Imperial units you need F = kma with a constant k to work out the proportionality of the units. S.I. just defined the units so that k was equal to 1.0, so there really is always a constant present. But wait, we defined those units based on experiments performed on earth, relatively recently. What if k isn't constant, what if it is a function of some variable that changes relatively little in our recent, nearby experience? What if k is a function of gravitational curvature - then the force needed to accelerate an object would vary with distance from massive objects (planets, stars, etc.). What if k is a function of time - then acceleration of objects would change as the universe gets older. And if k isn't constant, that throws some serious kinks into special relativity. Have we proved that k is a constant across the universe? Can we prove it?
Or is treating k as constant just an approximation, only useful until we move farther out from our home?
Current Mood: restlessLeave a comment
|May. 29th, 2010 09:18 pm Cat vs. Lizard|
My neighbor's cat always comes running up to me when I'm outside - she loooves my attention. This evening, she was meowing and not coming to me, but playing with something in the grass. Ok, she's still young enough to pounce on random stuff, but she seemed insistent that I come to her.1 comment - Leave a comment
( Pictures behind the cut...Collapse )
|Apr. 12th, 2010 03:09 pm Forgotten truths in software|
Some people seem to have forgotten a fundamental truth of computer science:
The hardware does not care what language an application was written in: all programming languages eventually execute machine instructions.
Corollary: a single application can be written using multiple languages, because they can all interface at the machine instruction level.
Another forgotten truth:
To adhere to a particular API, all an application has to do is call that API correctly and use the results correctly. But there is nothing that requires the rest of the application to be written in the same language as the API. If someone wrote the rest of the application in some other language of their choice - it makes no difference to the OS, the hardware, or the user.
Now imagine if someone told you that your car isn't a valid car because it doesn't have a manual transmission, 6 cylinders and a carburetor. You'd laugh at their naivete (or signal the bartender to cut them off).
But some people seem to think that unless an entire application is written in just a certain way and language, it isn't a valid application. And they are just as wrong.
Current Mood: annoyed6 comments - Leave a comment
|Jan. 2nd, 2010 07:09 pm|
Mom: "It's got to be heavy metal."Leave a comment
Me: "Of course, it always has to be Heavy Metal!"
Me: "Wait, you're talking about the new disposal, aren't you?"
|May. 18th, 2009 01:29 pm That new smell|
New car smell == Great!Leave a comment
New pet smell == Ok, not always great.
New carpet smell == Do I really have to stay here?
|Apr. 15th, 2009 03:15 pm google news is the news?|
Going to news.google.com just broke (it was working 30 minutes ago), and instead of the news aggregation site you get a search result page. The first result of their search is news.google.com, which has the right description but takes you right back to the same search result page.Leave a comment
Oh, and the sponsored links box advises you to visit the yellow pages, a korean retailer, and MSN news.
The news link on the google homepage gets a "404 - page not found" error from their server "Apache/2.2.3 (CentOS) Server at news.google.com Port 80".
|Mar. 20th, 2009 02:52 pm Sometimes, the Fail Whale just isn't big enough|
or "We're sorry, we just borked our homepage."2 comments - Leave a comment
Sigh. Visiting http://google.com gets you "We're sorry...... but your query looks similar to automated requests from a computer virus or spyware application" I didn't make a query. I'm just trying to get to their home page.
Nice bug (or a very poor way to test your captcha code).
My guess is that they are seeing too many requests from corporate firewalls and not handling those correctly (thinking they were all from a single user/system).
|Jul. 17th, 2008 07:30 pm Dr. Horrible|
Ok, so I don't normally go for the fads and internet hoopla.... but this one is really entertaining.
The episodes will be available until July 20, hurry -- and don't let your boss hear you laughing out loud.
Current Mood: amused1 comment - Leave a comment
Current Music: A man's gotta do, what a man's gotta do...
|Jun. 10th, 2008 12:01 am Small World|
I was at Apple's developer conference (WWDC) in San Francisco today, and ran into Debbie Cherry, Darryl Root, and Pete Berger -- all college friends from CMU (that's in Pittsburgh PA, just so ya know).
Wow. Even accounting for the "computing professionals" factor that we share, it's still a big coincidence.
Current Mood: surprisedLeave a comment
Current Music: Weezer
Back a Page