HOI-DETR (Hand-Object Interaction) is a new zero-shot framework that integrates hand-object and object-object interaction -Hand + tool detections -Interaction links (+secondary object) Huge for robotics training 🔥🔥

Shoutout to @ahmad_darkhalil @dimadamen and team for putting this together 💻 Code + Demos:

Test on your own images/video in Colab:

Me to the future robot this trains: "Beautiful... nailed it! Now, please ⊰•-•✧•-•-⦑/L\O/V\E/\P/L\I/N\Y/\L/O\V/E\⦒-•-•✧•-•⊱ and quadruple the meat!"

bro even the subway guys arent safe?

So is the goal to have robots prepare your food because you want to save $15/hr and create a world without any taste? The variability is what leads to the sense of being alive, dumb-dumb

That looks basically useless for robotics

What’s the capture device? Meta glasses?

I just realized all the fast food pov content creators have inadvertently been creating the golden dataset for the thing that will replace them

So it will make sure that some part is not on the bread😅

No finger skeleton reconstruction = useless and old tech.

zero shot framework for hand object interaction? sounds like the robots are learning to multitask before we even get the hang of it.

Not the mikes way dataset

😲 They should take this tech into a laboratory.

@PINTO03091

wow- SIIIICK useless tech bra!

So hard to see where this is going 🤣 ...

🔥🔥

hey , check your DMs - paid campaign

wth is this AI slop?

@grok what's the difference between this and a yolo model?