Picking Out The Object
from the objectification-in-a-good-way dept
MIT’s Tech Review has an article about a few different systems that are coming out in the near future that can take live video feeds and automatically pick certain objects out. This can be useful in a variety of ways. It can be used to cut down bandwidth requirements – as it will only be necessary to transmit movements among objects instead of the “background image”. It can also be used for security purposes – alerting guards (for instance) if a person enters a restricted area, or leaves a briefcase behind.
Comments on “Picking Out The Object”
This is new?
I thought CODECs had been doing this for years! I remember early AVI CODECs increasing framerate by only updating the parts of the stream that had changed. Guess I could be worng, though. Sure way to find out would be to post this message on Slashdot! You get flamed there for saying the sky is blue!
Re: This is new?
Yeah, I believe that it’s been possible to only send the part of the stream that moved for a long time. However, that’s slightly different than having the system actually pick out an object and only watch that object to see if it moves. So, while the old system still watches the entire images to only update the parts that move, this narrows what the system is even watching, to only pick up on if the important stuff moves.