NeatTrix

This is a little off topic, but it’s so damn cool I had to include it.  Where it is relevant is computer interface.  The Curious Case of Benjamin Button effects designers, were really the first to test the facial surface capture technology used in the production.  The video does a much better job that I do at explaining it, but for the first hour of the film, the character of Benjamin is played by various actors with a CG head modeled on Brad Pitt’s controlled, without the use of animator, by Pitt’s own performance.  Emotion capture, as they called it, is a novel computer interface.  An impractical one at the moment perhaps, but there is no denying this technology is remarkable.  Video is a bit long (17 min, but fascinating if you have the time)

We’ve looked at brain interfaces before (Emotiv Headset). This article looks at a different technique.

If you’ve ever happened to cover the end of a small flashlight with your thumb in a darkened room, you’ve probably noticed that your skin is remarkably transparent to red light. You may have also noticed a periodic darkening of your glowing thumb in synchrony with your heat beat. In broad strokes, that’s the principal by which the finger pulse monitors (plethysmographs) work. It works even better at near infrared wavelengths. Oxygenated blood and deoxygenated blood have maximum absorbances at different wavelengths. By alternating wavelengths, measuring the absorbance, and doing some math,  you can calculate the oxygen saturation of the blood. And that’s a pulse oximeter.

Researchers have long used blood flow in specific areas of the brain as an indicator of brain activity in that region. The principle is sound: increased neural activity requires lots of energy and oxygen, and that results in greater local blood flow. Researchers used to measure blood flow with magnetic sensors, but now use functional magnetic resonance imaging (fMRI) and similar imaging techniques, some of which measure glucose uptake more directly.

OK, now combine the two ideas and you might come up with something like Tom Chau, et al did at the University of Toronto. They used IR emitters and detectors to decode subjects’ preference for one beverage or another with 80% accuracy. It’s a start.

For more information:
www.sciencedaily.com/releases/2009/02/090210092730.htm
www.physorg.com/news153472589.html
www.crunchgear.com/2009/02/10/new-brain-scan-technique-uses-ir-detecting-headband
www.bioe.psu.edu/NMR/pdf_files/2005/enrico.pdf
www.bcs.rochester.edu/people/aslin/pdfs/Aslin_Mehler2005.pdf

Sharing more in common with the Microsoft Surface than the Minority Report-style waving your hands around in the air, I think this is much closer to the ideal of human computer interaction, baring any kind of neural uplink.  Direct manipulation with very little effort.  The keyboard is the only thing that troubles me.  This is 21st century technology being controlled, in part, by a 19th century data input device.  Sure, the materials used to make it are different, but the action of tying in words one letter at a time, one finger at a time is the same, with a pattern of letters layout out in order to slow the user down so the striking hammers won’t stick together seems silly to say the least.

Since we have a burgeoning thread of articles about this sort of thing, here’s yet another gestural display interface. iPoint 3D promises display interaction without a data glove. Two firewire-interfaced cameras feed live video to a dedicated processor which then tracks the motions of  hands and fingers in real time and relays the information to a computer. See www.sciencedaily.com/releases/2009/02/090220075312.htm and www.hhi.fraunhofer.de/en/departments/interactive-media-human-factors/overview/ipoint-presenter/ for more information.

I had a lazy eye as a child and refused to do the exercises that were supposed to train my eyes to work as a team again.  Because of my eyes inability to work together, my brain lost the ability to process both images at once.  I predominantly use my left eye to view the world.  My right eye is functional, if you call 20/200 vision functional, but I don’t look through it generally speaking.  About the only downside of all this is most 3D related techniques don’t work for me.  Anything that requires the normal processing of images from the left and the right eye is most likely lost on me.  Magic eye is still a mystery to me.  As is the following technology.

In fact, the only 3D technology I’ve encountered that actually does work for me are the glasses which flicker the left and right eye alternatively transparent and opaque.  By ‘forcing’ my eyes to work alone, one after the other, I get the effect that everyone else does.

From the MIT Media Lab, Siftables are both a computer interface and distributed computers in their own right.  They allow you to get literally hands on with your data.  The word and math applications I think are really interesting and have a lot of obvious applications in schools.

Originally developed at NASA (www.nasa.gov/centers/ames/news/releases/2004/subvocal/subvocal.html), subvocal speech recognition allows human-computer control through non-voiced speech recognition. Electrodes placed near the larynx on the throat pick up minuscule (< 1 millivolt) potential changes associated with speaking words without sound. For example, reading silently. That is, thinking that you’re going to say something you don’t actually voice. Evidently, bioelectrical signals similar to EMG are transmitted even though you don’t intend to say anything aloud.

The Ambient Corporation (www.theaudeo.com) has commercialized this technology, recently making the first voiceless phone call (edageek.com/2008/02/27/msp430-microcontroller).  Below is a video demo using this technology to control a wheelchair. The biggest drawback of this, EMG, or any electrode-based human-computer interfacing are those damned electrodes. Daily use irritates the skin, and then there’s all the motion artifact, not to mention that you have wires hanging off your face.  Implantable electrodes? Feeling Borg-ish?

When I saw the film, Minority Report, with Tom Cruise, I was completely intrigued by the data glove and gestural interface depicted in the film, and even worked on developing an interface with some of the same ideas in mind. Well, I’ve been scooped again. Check this out:

I am so jealous…

If you are like me, then you have about 137 wall warts and no way to plug them in. Newer wall warts tend to be more “outlet-friendly”, but many still will not allow use of adjacent outlets when plugged in. Grrrr!

To help solve this dilemma I buy short power cords to move the wall warts off the strip. Maybe you already do this. A good, reasonably inexpensive source for these is Jameco (about $2.50). They also sell a twofer cable for about $6.50.

www.jameco.com/webapp/wcs/stores/servlet/ProductDisplay?langId=-1&storeId=10001&catalogId=10001&productId=199778&

www.jameco.com/webapp/wcs/stores/servlet/ProductDisplay?jameco_page=42&langId=-1&productId=208347&catalogId=10001&freeText=208347&storeId=10001&search_type=all&ddkey=http:StoreCatalogDrillDownView

Michael Linderman and university researchers (www.norconnect.com) are exploring the use of foream EMG to recognize handwriting. The idea would be to interact with computers by interpreting the forearm muscle motions associated with writing. This is in the very early stages. Polygraph amplifiers are used to get the EMG signals. Everything needs to be miniaturized. Linderman’s group has received federal SBIR funding to develop this technology. See the following article for more information: www.sciam.com/article.cfm?id=handwriting-recognition

Happy Arbitrary Time Day!

Unix time will reach 1234567890 February 13th at 5:31pm EST. For the uninitiated, there are several ways of storing the date and time information on a computer, one of them is called Unix time, also called POSIX time. The Unix Epoch was January 1, 1970 at midnight. Unix time is represented as the number of seconds that have passed since that time. It will reach the marginally interesting count of 1234567890. I will try to post this as close to the time as possible.

Self-leveling technology.  A neat idea. Probably not cost-effective for restaurant tables, but still pretty cool.  After watching their ridiculous ad on their website: flat.net.au you’d think they were selling wine, or a time-share in Monte Carlo, but no, it’s an automatic, hydrolic, table leveler. Yay!

Actually they claim it’s for any application where stability and being level are important.  Depending on what your application was and knowing how much it costs this could be a very useful technology.  On the other hand, unscrewing the foot of the table out a bit to level it, seems like the much simpler and cheaper solution.

Video after the jump