Yoni Kasten
@yoni_kasten
Followers
305
Following
116
Statuses
86
Sr. Research Scientist at NVIDIA Research
Joined June 2015
🚀 Our code and data are now live! Explore them here: 📌🔥 #4DTracking
Excited to present TracksTo4D at NeurIPS 2024! Our method enables efficient reconstruction of dynamic 3D point clouds and camera motion from casual videos—all in a single forward pass! This work is a collaboration with Wuyue Lu and @HaggaiMaron. More details in the thread! 1/9
1
40
221
RT @sara_oblak: Reconstructing dynamic scenes from video inputs is challenging due to the sparse nature of inputs, both in time and space.…
0
28
0
9/9 Our code is coming soon! For more details, visit our project page:
0
0
7
RT @YoadTewel: 🚀 Excited to release the code and demo for ConsiStory, our #SIGGRAPH2024 paper! No fine-tuning needed — just fast, subject-c…
0
33
0
RT @ChenTessler: Excited to share our latest work! 🤩 Masked Mimic 🥷: Unified Physics-Based Character Control Through Masked Motion Inpai…
0
91
0
RT @YoadTewel: I'm going to present ConsiStory📖 at #SIGGRAPH2024 this Monday @ 2pm! If you're around this week, DM me if you want to chat!…
0
20
0
RT @YarivLior: Thrilled to share Mosaic-SDF (M-SDF), a simple 3D representation suitable for 3D generative models! Check out more results o…
0
74
0
Today at 10:45 am, we will present SDS-Complete at #NeurIPS2023, poster #110. Come say hi!
Excited to present SDS-Complete at NeurIPS 2023 this week! A test time optimization method for point cloud completion with pre-trained text-to-image Diffusion Models. This work was done together with @OhadRahamim and @GalChechik. Project page:
0
0
8
RT @DanahYatim: Our code is now public!!🥳🎉 Webpage: GitHub: Can't wait to see all the cool…
0
30
0
Today at 10:45 am, I’ll present our #NeurIPS2023 paper "SceneScape: Text-Driven Consistent Scene Generation" at the Great Hall & Hall B1+B2 (level 1), poster #119. Come say hi!
Excited to present "SceneScape"! Generate long-term walkthroughs in imaginary scenes, such as ice castles, caves, or spaceships, just from an input text prompt! [1/3] @amit_abe
@RafailFridman
@yoni_kasten
0
1
15
RT @maggie_albrecht: #NVIDIAResearch collaborates with leading universities to advance #generativeAI, robotics, & natural sciences. At #Neu…
0
3
0
7/7 It is possible to control the completion by using different text descriptors for the same input point cloud.
0
0
0