Notices

I'm amazed at the ammount of information is taken from the markers.

I think the markers hold only the ID of the model. The 3D model info is built into the arapp app (hard coded) IMHO. Just a demo, but very nice indeed.

Yep this is exactly it, in pseudocode:

Load Model 'X'
Load Model 'Y'

When marker #1 appears:
Get marker position
Get marker rotation
Apply these attributes to model 'X'


When marker #2 appears:
Get marker position
Get marker rotation
Apply these attributes to model 'Y'

I know all of you are going to laugh at me saying 'shame', but I don't understand how to use the application! I printed the pdf, but I don't know what I have to do with it!

Place the A4 marker sheet onto your desk.
Open the N900 camera shutter.
Close down camera app (That will probably start)
Open arapp app
Point the camera at the A4 sheet, hold nice and steady for a bit.
Wonder at the future of computing.

OK arapp is a technology demo, but it's pretty fab.

PS for anybody interested in AR, I can recommend the PS3 game 'Heavy Rain' it has an interesting character in it. Anyway I digress!

was trying to get it working but Arapp closes immediately after i start it :/ help anyone?

http://www.youtube.com/watch?v=G52wEqOSE_0

Erm, we could use the GPS integrated compass when walking. Could we please still get AR?

I am having the same error. A little deeper look into the terminal, showed me that it is a segmentation fault.

You don't even need to print marker out, to test it point camera directly onto your monitor. (don't throw your monitor from the desk trying make it flat, this is just for testing purposes ).

I'm really interested in this technology from many years. Flash guys are playing around AR since the library has been ported to AS some time ago.
I was trying to investigate many different recognition algorithms (some of ideas came from 60's) This technology has big potential IMHO. However if you truly and deeply digging into this area only one conclusion comes to your mind.

All of algorithms we dealing with these days are pretty much base on flat simple shapes. No matter what (Face recognition algorithms, traffic, move tracking, OCR, some 3D Google stuff, some are based on neural network etc etc) are based on 2D picture.

Mom Nature gave all of living beings pair of sensors called eyes or ears for reason. We worked it out how to do voice cancellation or noise reduction using 2 mics right? Even from programing point of view would be a lot more efficient to have 2 sources (2 eyes/cameras) to simplify algorithms and reduce calc. power needed. Apart from advantage that 2 cameras will give us, 3th dimension on the fly.

Before AR systems will not get 2 eyes this technology will be just a prototype for fun. Nothing serious will happen because 2d algorithms are over-killed. I am really surprise why still so many research is going with death end solution.

But imagine now. This is my dream idea so far... Future mobile device. Pair of glasses that you can see whole GUI interface on your hands. Manipulate on virtual objects. No display needed, no buttons, no keyboard. Just glasses, some mic and earphone and welcome to the Matrix Neo . It might sounds silly but with AR I can see it possible.

Remember this post and back here around 2020