Colton Stearns

Robot Action Models @ Luma AI

[Luma @LumaLabsAI]

Direction goes in. Cinema comes out.

Ray3.2 is here → https://t.co/TuAuFhZDBn

[Luma @LumaLabsAI]

Uni-1 is here! A new kind of model that thinks and generates pixels simultaneously. Less artificial. More intelligent.

[Luma @LumaLabsAI]

Try Luma today. - https://t.co/GdPpsPAb8z

[Luma @LumaLabsAI]

Introducing Luma Agents. Creative agents that make you prolific. You set the direction. They build with you, seeing what you see and helping teams explore further, iterate faster, and watch ideas multiply.

[Luma @LumaLabsAI]

Introducing Uni-1, Luma’s first unified understanding and generation model, our next step on the path towards unified general intelligence.
https://t.co/QjdrnYoWe5

[Luma @LumaLabsAI]

This is Ray3. The world’s first reasoning video model, and the first to generate studio-grade HDR. Now with an all-new Draft Mode for rapid iteration in creative workflows, and state of the art physics and consistency. Available now for free in Dream Machine.

[Jihyeon Je @JihyeonJe]

Symmetries are everywhere — from butterfly’s wings to Greek temples. But detecting them in noisy data? That’s a challenge. 🦋🏛
Our #SIGGRAPHAsia2024 paper, Robust Symmetry Detection via Riemannian Langevin Dynamics, tackles this: https://t.co/zVkv4LZv6N
🧵(1/n)

[Colton Stearns @Stearns2Colton]

@WayneINR @AdamWHarley @mikacuy @fedassa @fpgDubost @GordonWetzstein @GuibasLeonidas We're interested in this! We primarily focused on novel view synthesis in this work. I think there's great promise for more explicit 3d tasks, but will likely have to tweak rigidity losses + incorporate stronger surface-like constraints.

[Colton Stearns @Stearns2Colton]

Video to 4D is an incredibly challenging task! Excited to announce that Dynamic Gaussian Marbles was accepted to #SIGGRAPHAsia2024 with @AdamWHarley @mikacuy @fedassa @fpgDubost @GordonWetzstein @GuibasLeonidas. Check out our webpage and code!
https://t.co/rLz8cRfU1B

[Colton Stearns @Stearns2Colton]

@_ilya_c @AdamWHarley @mikacuy @fedassa @fpgDubost @GordonWetzstein @GuibasLeonidas Really depends on the video. Sometimes the other losses (photometric, LPIPs, depth, and segmentation) can guide it to appropriate correspondence. Other times, correspondence fails and you end up with a jumble of Gaussians.

[Colton Stearns @Stearns2Colton]

Video to 4D is an incredibly challenging task! Excited to announce that Dynamic Gaussian Marbles was accepted to #SIGGRAPHAsia2024 with @AdamWHarley @mikacuy @fedassa @fpgDubost @GordonWetzstein @GuibasLeonidas. Check out our webpage and code!
https://t.co/rLz8cRfU1B

[Adam W. Harley @AdamWHarley]

Very excited to speak tomorrow at the CV4Animals workshop @CVPR, where I will be talking about the huge team effort from my past year at Stanford:

"Building a Foundation Model of Motion: Tracking Anything at Any Granularity"

Talk details: https://t.co/WVrgW0hzgJ
#CVPR2024

[Colton Stearns @Stearns2Colton]

I'm excited to present CurveCloudNet this Friday afternoon @CVPR with my collaborators @paschalidoud_1 @davrempe @jjpark3D @JiatengLiu @GuibasLeonidas. It's been a long time coming, showing an exciting new way to think about point cloud perception! https://t.co/zsvYmAkx8D