A gesture recognition system to transition autonomously through vocational tasks for individuals with cognitive impairments.

Two adults with intellectual disability tried a Kinect game at work. The camera watched their hands and gave step cues for packaging tasks.

When the worker missed a step, the screen flashed green and showed the next move. No staff had to speak.

Correct steps jumped from a large share to a large share when the Kinect was on. Scores dropped again when it was turned off.

The workers kept the gains after only two sessions.

Diemer et al. (2023) got the same lift using short phone videos instead of a motion camera. Same goal, cheaper tool.

Bigby et al. (2009) and Spanoudis et al. (2011) used PDAs to prompt cooking and task boxes. Their fade-out logic matches the Kinect, but the device was smaller.

Spriggs et al. (2015) let parents film the teen himself doing the task. The self-modeling clips worked faster than generic cues, showing faces matter.

You can trade the pricey Kinect for a tablet and still get independence. Start with video prompts; add self-modeling if the learner stalls. Fade the clips as you would any prompt. The tech changes, the teaching rules stay the same.