Object recognition.
Then “action recognition”. I mean, what the objects are doing. To each other and such.
Then you have a narration machine. Which could be nice.
Object recognition.
Then “action recognition”. I mean, what the objects are doing. To each other and such.
Then you have a narration machine. Which could be nice.
over_clox@lemmy.world 1 year ago
Very related, not my work though…
Self Aware Lara Croft
This seems like something of an ideal effort, but it also took lots of human work to even prepare the AI model, and many many runs to refine it.
We’re not even close to realtime object recognition at this point. Delayed recognition, yes. Realtime, no.