Prototype Camera Describes a Picture’s ContentBy: Chris Gabbard - April 26, 2012
Imagine a camera that doesn’t take a picture; it gives you a description of the scene you are viewing. Doesn’t sound too appealing, you say. But a picture is worth a thousand words, you say.
Well, imagine a camera taking and picture AND giving you an instant description of the scene. Now the uses of this technology are starting to make sense. In today’s snap happy society, having those descriptions and being able to search for them would make a huge difference in cataloguing and retracing all of your fondest memories.
Imagine having all the pertinent information about a photo attached to each one automatically. Information like who is in the picture, where it is taken, and what everyone is doing would be key in keeping a clean photo album.
That’s what Matt Richardson is working on with his Descriptive Camera. Right now it is only a prototype, a bulky prototype at that. But it is an option to explore.
The core technology used to create the descriptive camera is Amazon’s Mechanical Turk API. According to Richardson, “It allows a developer to submit Human Intelligence Tasks (HITs) for workers on the internet to complete. The developer sets the guidelines for each task and designs the interface for the worker to submit their results. The developer also sets the price they’re willing to pay for the successful completion of each task. An approval and reputation system ensures that workers are incented (sic) to deliver acceptable results.”
So basically what he is doing is sending a picture to a worker who is writing out a description of the picture and sending the information back. Wait. That isn’t impressive at all. I could have come up with that, and I would have had the good sense to not make a prototype of such a moronic idea.
You can even set the camera to “accomplice mode”, sending an instant message to your friend asking them to create a picture description for you. Like, here catalogue my photo album for me…
Sending the photo to Mechanical Turk costs $1.25 per photo and takes 3 to 6 minutes for the information to return. This has got to be the stupidest idea I have ever heard of. It doesn’t even take three minutes to write your own description of a photo, a feature most photo-sharing and cataloguing programs accomplish. And it doesn’t cost $100 to complete an album.
Find a way to use computer AI to make a description and you might have accomplished something.