A combination of hardware and software that would keep track of all your belongs inside your home. If you wanted to find (say) your scissors, you would just search for them on the system and it would tell you where they were last seen.

There have been a spate of systems that try to solve this problem by attaching a piece of hardware to each object to be tracked, but this would be impractical and expensive for more than a few objects.

OrsonObject would be pervasive and, once set up, would automatically keep track of everything by tracking objects by their appearance. To set up the system, all you would need to do is install a bunch of tiny cameras around your home such that they had a pretty good view of almost everywhere. This is not expensive or as much of a hassle as it sounds- there are systems like VueZone with tiny wireless cameras that you stick onto the wall with double sided tape.

I am not a big fan of the VueZone system specifically, but it shows that this technology is nearly ready. Since the OrsonObjects system doesn’t need to stream video, it would be possible to make the cameras even smaller and cheaper.

The cameras would automatically take pictures whenever anything happened in front of them- typically you moving stuff. These images would get uploaded to cloud-based servers in the background. These servers would use continuous recognition to find and any objects and track their movement throughout the space.

This is hard, but possible (or almost possible). People have made a lot of progress in automated object recognizers lately. Since the processing is all done off-line, you could bring to bear huge, cheap computational resources to do the recognition and tracking.

The user interface to the system would look a lot like Piccassa’s face view, showing a list of all recognized objects. You could sort the list by things like object size, most recently moved, or most often moved. You’d click on the object you are looking for and it would bring up the last time that object was moved – directing you to the object’s current location.

These system could be extended to recognize objects that are often used together and group them logically into sets. It could keep track of most used objects and predict when you might need an object based on what other objects you are currently gathering (he has the coffee grinder…. and the sugar… which usually means that next he will need the filters also- which I last saw in the top drawer) and suggest their location to you when it seams like you are looking.

Looking ahead into the future, the system could be expanded to also manage the location of objects for you. You’d have a piece of furniture that looked like a large bureau. It would have a large, easily accessible shelf top in front. Anytime you were done using an object, you would drop it on the shelf and the system would automatically store it inside. Objects would be stored much very efficiently than you could in normal drawers and cabinets since the system could put every object where it fit best. Objects would be retrieved in seconds any time you needed them. This would save users both time and space.

UPDATE 5/24/15-http://www.wired.com/2015/05/next-roomba-may-recognize-crap/

UPDATE 12/31/2016 – http://www.theverge.com/2016/12/30/14129576/microsoft-hololens-key-finding-milk-tracking-patent