Will positional head tracking be activated in the dev kit?

"jwilkins" wrote:
Hmm, their key finding seems to be that an accelerometer can compensate for only having a single camera because it lets you resolve ambiguity by providing a proper scale for what is seen in the video. (Sorry if I seem to be using the forum instead of Google, but honestly I did not even know to use IMU as a keyword before, since I'm relatively new to this).

The important thing is not why they were fusing them, but how.

And while searching for "video accelerometer fusion", it would seem that ONCE AGAIN apple was awarded a patent for something people have been doing for quite awhile already (using accelerometer and gyro data for digital video stabilization). Getting patents on the color black and rounded corners was bad enough, but now this? The patent system is badly broken...
http://www.macnn.com/articles/11/09/29/utilizing.accelerometer.and.gyroscope/

I'm a little confused by your distinction between why and how. Pretty sure I just rephrased their paper title to make sure I understood what they were doing.

I'll do a proper literature search after I'm done writing this conference paper that is due today... promise.

Ugh. More insane patents. Just great.

There's a long list of patents that clearly should never have been granted, and on top of that there's a reasonably large body of evidence showing that software patents in particular are highly damaging to everyone involved. - So much so in fact, that it might be reasonable to argue for making it impossible to patent software algorithms due to just how destructive they are. (For one thing, some estimates show the cost of litigation related to software patents is something like 5 times their actual value in any commercial applications... - You know something's badly wrong when a system designed so that people can make money from their inventions soaks up substantially more money in legal costs than the inventions are even worth in the first place. - It stops being a system to encourage innovation, and starts being one that cripples it instead.)

Speaking of crippling innovation, the steam engine condenser had a patent on it for more than 50 years, which, because it so drastically affected the efficiency of a steam engine, and no viable alternatives existed, basically crippled any improvements to steam engine technology for the duration of it's existence.

History seems to show that the patent system in particular can often have the opposite effect of it's officially stated purpose...

Hey, geekmaster, you quoted my question about CV libraries but you didn't give a recommendation. Did you forget or did I miss something?

"jwilkins" wrote:
Hey, geekmaster, you quoted my question about CV libraries but you didn't give a recommendation. Did you forget or did I miss something?

You probably know more about this stuff than I do...

A popular CV library is OpenCV:
http://opencv.willowgarage.com/wiki/

But there are other options as well, depending on your needs.

SimpleCV - Computer Vision platform using Python:
http://www.simplecv.org/

OpenVX - Hardware acceleration API for Computer Vision applications and libraries:
http://www.khronos.org/openvx

STAIR Vision Library:
http://ai.stanford.edu/~sgould/svl/

Fast CV (for mobile devices):
https://developer.qualcomm.com/mobile-development/mobile-technologies/computer-vision-fastcv

I have very little personal experience with Computer Vision libraries, so I can only give recommendations based various research papers and whatever Google can find. I personally prefer small and simple solutions when they are able to do the task I need for my current project.

lol, I was just wondering if you did have a rec, but just forgot to add it :)

If you didn't have one then it is OK to say so. I've made a note of how helpful you are (after seeing so many helpful info-dumps from you in many threads) and will try to not use you as a personal Google anymore :)

googlemaster...

I've been using OpenCV for my SLAM project over the past few months. Of the major CV libraries, it's certainly got the broadest support for newer feature detectors and descriptors. You'll still run into "gotchas" on some of the newer code, but it's generally pretty solid. And the low-level stuff that Intel did is very mature.

have you seen this topic ?

viewtopic.php?f=25&t=787&p=11215&hilit=radio#p11215
and this video :
https://www.youtube.com/watch?feature=player_embedded&v=mYyFUQbWC1E

"Tgaud" wrote:
have you seen this topic ?

viewtopic.php?f=25&t=787&p=11215&hilit=radio#p11215
and this video :
https://www.youtube.com/watch?feature=player_embedded&v=mYyFUQbWC1E

That will be a cool device when it becomes available to us at a price we can afford. But for now, the only affordable and viable options are attaching a Razer Hydra controller to your head, or using an optical tracking method. Perhaps a software solution will become available at some point too (limited to seated or standing in place, and perhaps gesture-based, but still useful for a "no-additional cost" gaming solution).