Author Topic: Tongue Tracking  (Read 6279 times)

Offline Rohan

  • *
  • Posts: 4
  • Karma: +0/-0
    • View Profile
on: July 25, 2012, 12:50:38 AM
Hi,

I am doing a project to help people who have  sustained high level spinal cord injuries, it involves using the tongue to operate a computer. I have been testing the idea through using reflective tape on the tongue, the cursor tracking works well but the range of mobility of the tongue is too small to navigate across the entire screen. Is there anyway to increase the input gain through either the Wiimote whiteboard or a standard windows program? Also how do you stop the movement of the cursor being inverted, is there any programs instead of Wiimote whiteboard which can do this?

Thanks in advance,

Rohan



Offline zainey

  • *
  • Posts: 64
  • Karma: +2/-0
    • View Profile
Reply #1 on: July 25, 2012, 02:58:20 AM
Hi,

To amplify the 'gain' of the wiimote, you may consider placing a telescopic type lens in front of the wiimote camera.

A telescopic lens - like a rifle scope - would expand the effects of small lateral movements of a tongue, say to the viewing area of the wiimote. You may have to reduce the apparent 'blob' size as seen by the wiimote, as it seems to 'go blank' when viewing large 'blobs' up close.

Use the Freetrack software, in single LED point mode, to help in visualising what the wiimote sees, and to make adjustments for the perceived sensitivity of the wiimote IR camera as well.

You may also need to experiment with different narrow field of view lenses to see which kind works best. Even a compact CCTV type lens may do the trick.

Alignment is fairly straighforward. Line up the external IR LED with the bare wiimote to the centre coordinate (512, 384). Once the IR LED and wiimote are inline, then place the telescopic lens in the light path and move the lens around so the IR dot is re-aligned to the centre. Then fix the lens in place to use properly.

Hopefully that puts you on the right track. p

Thanks.




Offline Rohan

  • *
  • Posts: 4
  • Karma: +0/-0
    • View Profile
Reply #2 on: July 25, 2012, 08:00:37 AM
Thanks zainey, it's a pretty interesting approach that you have suggested, I'm going to give it a go now and I'll let you know what eventuates.

Thanks again!!



Offline Rohan

  • *
  • Posts: 4
  • Karma: +0/-0
    • View Profile
Reply #3 on: July 25, 2012, 11:11:59 AM
The Free track program has come in pretty handy, pretty much solving the gain problem with the autopan setting. However the autopan setting does not start working until the IR blob is a certain distance away from the center, does anyone know how to reduce this so it would work with the tongue? Because it cannot quite move far enough to trigger it.

Otherwise I will try and find a lens to use, as zainey suggested



Offline zainey

  • *
  • Posts: 64
  • Karma: +2/-0
    • View Profile
Reply #4 on: July 25, 2012, 09:33:36 PM
Hi,

The 'curves' settings in FreeTrack may help improve the lack of response reported around the centre position.

The curves menu feature can be used to amplify the apparent motion of the IR blob and thus make up for some of the limitations in the field of view or the actual range of IR dot motion.

Its not something I ordinarily tinker with, so I hope it helps with your application.

Thanks.



Offline Rohan

  • *
  • Posts: 4
  • Karma: +0/-0
    • View Profile
Reply #5 on: July 28, 2012, 11:13:20 AM
Hi,

Thanks for the suggestion ! I had a quick read of the Free track handbook, and if you right click on the curve you can select to have a small deadzone, this amplifies small movements about the center point, this works like a treat. Just have to make sure when you click on autopan that your IR source is in the field of view, or else it wont work and do stupid things.