Feasibility of sorting a basket of clothes in the real world

Question

I'm wondering how feasible it is to create a machine that can separate clothing from a basket.

At the most basic level it would distinguish between tops, pants, button downs and socks

Programmatically, I'd image this would require training a neural network to recognize these items, but in real time it becomes exponentially difficult to do this in a small space at a fast rate:

1. pick up an item
2. lay it in such a way that is recognizable
3. deduce whether it is a top, button down, etc.
4. sort it accordingly

If this sounds ridiculous please let me know...

If it is possible :

would this be based on some sort of computer vision? or only a well trained neural network?

Any insight is much appreciated!

Answer 1

Peter Abbeel does work in deep learning for robotics, and one of the projects they've tackled is manipulating clothes. Here's a video from 2011 of their robot folding laundry (one piece at a time).

There are also companies attempting to market this; seven dreamers makes the Laundroid and Foldimate claims that it will start taking pre-orders in 2017.