[visionlist] About help on eye-tracking of head free non-human primates

[Jeff Mulligan]

On 7/17/18 4:45 AM, Gabriel Diaz wrote:
> /3. Are there multi-camera eye-tracker systems that we can set-up 
> fromdifferent view points so that one or more can always have a clear 
> viewof the animal?/
>
>     Not that I've seen. I have discussed building something like this
>     before with colleagues.  That would be a feat of engineering and
>     software development that requires a very firm grasp of multiview
>     geometry.  That is a multi-year project.
>
>
The SmartEye system uses multiple cameras.  The have built-in head and 
facial feature models that can often track a new subject reasonably 
well, but for best performance the models need to be tuned for each 
subject, which is a procedure that is done from within their software.  
It might be possible to get this to work with a macaque, but you would 
need to talk to their engineers.

I don't have much first-hand experience with this system, but gaze 
accuracy is to a first approximation proportional to the number of 
pixels devoted to the eye, and so any system where the camera captures a 
large head box will necessarily suffer (unless you have a camera with a 
LOT of pixels).  So it depends on what your accuracy requirements are.

good luck!

-Jeff


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://visionscience.com/pipermail/visionlist_visionscience.com/attachments/20180717/a39d2b28/attachment.html>