A key thing limiting reasoning about 3D human-object interaction in images is the lack of training data. In PICO, we propose a novel way to label human-object contacts in images and build a new dataset with dense, vertex-level, contacts on the humans and objects. We then infer the 3D object shape and pose, human shape and pose, and use contacts in an optimization method to solve for a consistent 3D configuration.
I’d love to humbly introduce you to something we believe is going to be a game changer for both VR and VFX.
When I first tried it I couldn’t take a headset off for like 10 mins. The hair, the dress, the movements — your brain just can’t stop being mesmerized.
Tech thread 🔽
@tszzl The predictable hype and commercialization of the concept proves that most of tech Twitter is just sheeple trying to stay 'on trend' so they can get 'engagement'.
I've experienced people bend themselves into pretzels trying to find a reason to not back a black founder. It's so stupid I would think I'm gaslighting myself into believing leftist propaganda but I've literally watched it happened and seen the alternative.
** VC TIP **
When you outsource your technical diligence to a 3rd party, you are also potentially adding gender or racial bias.
The technical person finds issues with the minority deal and hides his racist feelings in negative feedback and technical jargon. Sick.
I'm so unimpressed with generative images it makes me sad. When consistency and predictability can't be delivered reliably- it's functionally useless for world building.
People screw themselves by glamorizing what they're doing like it's going to change EVERYTHING, but usually they're just gaslighting themselves into doing nothing of value. Plumbers don't think like this. But plumbers consistently make money and deliver value.