Of course it's possible. Has coding taught you nothing?
Do any of you think its possible to use deep learning to learn shapes of players, so we can determine what are players or not solely on the image output, then create an aimbot tuned to specific games using deep learning tools? I'm not well versed in deep learning, I was just wondering if it was possible, and the time frame that would be possible for one person (at minimum would be an open source project) to create such a project. (Like just complete ballparking. months, years decades?)
Deep learning is a subject i'm possibly interested in learning.
Of course it's possible. Has coding taught you nothing?
I know its possible, but i'm more wondering I can do it or if its going to take an entire development team at google. Also it might not be possible with current hardware limitations, such as scanning each frame for what might be a player, then checking if it was right to learn, then transcribing that to a database.
- - - Updated - - -
^^ I forgot to quote
Last edited by Gratin; 11-12-2015 at 02:58 AM.
You'd be better off emulating a model's position like the game does and using that.