Not sure if you are looking for a cooperation rather than an actual answer, but when developing with kinect OpenNI is the right starting point: www.openni.org/

The development with this SDK is quite easily affordable also from a student doing a thesis...

Thanks for your answers, its very interesting. Start the development with OpenNI seems simpler than the Microsoft SDK.

It's good to be aware of some, often not mentioned, Kinect's drawbacks. Look here:

Obdrzalek S, Kurillo G, Ofli F et al. (2012) Accuracy and robustness of kinect pose estimation in the context of coaching of elderly population. In: Proc IEEE Annual International Conference on Engineer-ing in Medicine and Biology Society (EMBC) pp. 1188-1193

I'm not sure if Kinect's capabilities should be your main concerns, since it actually does what it says pretty well and is not that hard to program.

Btw, Leap Motion might be a good alternative also, at least is worth checking what it has to offer for your project.

However, interfaces for people with intellectual disabilities sounds way more challenging. Nowadays you can track people's faces, movements, gestures and so on. But the real question is how do you interpret and use them.

Guilherme DeSouza at the University of Missouri has been doing quite a bit of work using Kinect to develop tools for individuals with disabilities:

http://web.missouri.edu/~desouzag/publications.php

Your goals sound wonderful! Good luck.

Carla

Thank you Carla.

Hi,

I have been developing user interfaces using the kinect and openni for some time now and I can agree to the comments above, yet you should consider that Apple has recently bought Primesene, the company behind the first Kinect and openni, and has obviously no amibition to continue the development on the kinect and openni.

The primesense devices are no more sold and there are no comments on the future of the openni SDK yet.

So if the existing SDK, technology and functionality is sufficient for your tasks you should be fine, but when starting off a new big project you should be aware that the future of this technology is not safe at the moment.

http://community.openni.org/openni/topics/future_of_openni

http://news.cnet.com/8301-13579_3-57613893-37/apples-primesense-buy-its-all-about-the-maps/

an addition:

the differences between the Microsoft SDK and openni are that:

the skeleton tracking of Microsoft SDK is clearly more stable and accurate than the one by openni. Openni has trouble with the orientation of people when turning away from the sensor, it always assumes people face towards the Kinect.

On the other hand, openNI is way more flexible and easier to program. You can

- access all data streams and information for all users that were detected.

- decide who is the active user of the system,

- use implemented gestures

- simple saving and loading of training video material

in my application, the restrictions in the Microsoft SDK made it impossible to use for my research inspite of the superiority of the tracking algorithm because of the closedness of the system.

Thanks, Benedikt.

The most obvious drawbacks of Kinect would be that it can only be used indoors and that sensors may interfere with each other (and any other significant IR source, if any). I'd go for OpenNI, it is way easier to program and offers enhanced functionality. Plus it is already integrated in systems like ROS. Indeed, OpenNI follows a programming philosophy similar to ROS: devices and program modules are modeled as nodes that interact with each other, conforming the so-called production chains.

There are alternatives to Kinect that you may want to consider: http://wiki.ipisoft.com/Depth_Sensors_Comparison

However, they all use the Primesensor, so if Apple has really bought the company and they decide to discontinue, this is it.

The alternative would be TOF cameras, but they are way more expensive (9000EUR last time I checked) and, in my opinion, less flexible.

I am working with Kinect for assisting the disable patient . Microsoft Kinect provides a robust hardware and the data achieved are quite stable. My work is presently undergoing trials at a disability clinic.

Well, my answer it's not directly related with development, however I think you could obtain information with some words; visión based interfaces and cerebral palsy users.

In my opinión, I think you can add to your work some important aspects related to usability factors. If you're thinking to develop user experience studies it's important how to measure performance, effectiveness, efficiency and satisfaction.

I'm working with Kinect and Leapmotion. You can créate and develop applications, however perhaps your purpose is improve the quality of life of humans.

My recommendation: this paper about

Usability of visión based interfaces, C, Manresa-Yee, E. Amengual, P. Ponsa, 2013

https://upcommons.upc.edu/e-prints/handle/2117/22366