ML researcher. Previously, @berkeley_ai @iitbombay.
Berkeley, CA
Joined June 2012
- Tweets34
- Following317
- Followers551
- Likes304
Friday 10:30AM Spotlight Poster#132
SAP3D is a system for 3D perception which improves as the number of views increases. Unposed few-view NVS/3D by test-time adaptation of diffusion NVS models.
[Poster alert - Friday morning #132]
Most of the 3D reconstruction and novel view synthesis systems take a fixed set of images as input. They can not improve when more unposed images are provided, even if they reveal new information about the 3D structure.
Thrilled to be presenting "Humans in 4D" at #ICCV.
We reconstruct humans in the wild with remarkable robustness and accuracy! That too, with a very simple transformerized architecture.
Thursday afternoon *Poster 76* in the South Foyer!
🌎 shubham-goel.github.io/4dhum…
Humans in 4D, w/ @goelshbhm @geopavlakos @jathushan @JitendraMalikCV
✨HMR2.0💃🏻 + better PHALP, it's super robust! Give it a try:
🌎shubham-goel.github.io/4dhum…
🤗huggingface.co/spaces/brjath…
*Thursday afternoon poster*
Happy to share that I’m PhD-done! Berkeley was such an enriching and cherishable experience, special thanks to my advisors @akanazawa and @JitendraMalikCV for supporting me throughout.
Next, super excited to join @Avataar, where we’ll continue pushing the boundaries of 3D tech.
Thanks @a16zGames for hosting an amazing event and inviting me to speak. I learnt so much!
Shubham Goel retweeted
Releasing 🌟nerfstudio🌟, a plug-and-play python library to easily create your own NeRFs!
@nerfstudioteam is a contributor friendly open-source repo with a realtime web viewer that makes it easy to make cool videos 📽️
github.com/nerfstudio-projec…
docs.nerf.studio #nerfstudio 1/
Shubham Goel retweeted
Happy to announce DreamFusion, our new method for Text-to-3D!
dreamfusion3d.github.io
We optimize a NeRF from scratch using a pretrained text-to-image diffusion model. No 3D data needed!
Joint work w/ the incredible team of @BenMildenhall @ajayj_ @jon_barron
#dreamfusion
Shubham Goel retweeted
🎊 New paper!
We train loss-conditional diffusion models of *neural network checkpoints* that learn to optimize.
w/ Radosavovic, Brooks, Efros, Malik
proj: wpeebles.com/Gpt
arxiv: arxiv.org/abs/2209.12892
code: github.com/wpeebles/G.pt
Shubham Goel retweeted
For recognition to graduate from 2D to 3D we need new tasks.
Omni3D is a large benchmark for 3D object detection in the wild with over 230k images & 3 million instances and draws from existing datasets.
Our model trained on Omni3D generalizes to new domains like AR captures
Shubham Goel retweeted
Vision datasets (i.e. ImageNet) are usually collected once for a fixed task. But how do we know the choice of camera intrinsics, tasks, etc. is a good one?
Our ICCV paper on “steerable datasets” addresses this problem and gets 'human-level' surface normal preds along the way (⅓)
Excited to share our new work "Differentiable Stereopsis" -- on reconstructing object shape and texture from *few* images, masks and *noisy* camera poses! w/ @georgiagkioxari and Jitendra Malik.
shubham-goel.github.io/ds/
arxiv.org/abs/2110.05472
Shubham Goel retweeted
Super excited to share something we have been working on for the last 1.5yrs. Check out our #RSS2021 paper on Rapid Motor Adaptation.
RMA allows a legged robot trained *fully* in simulation to *adapt* online to diverse real-world terrains in real-time!
ashish-kmr.github.io/rma-leg…
Shubham Goel retweeted
#CVPR2021 Check out Motion Programs, a neuro-symbolic representation that expresses human skeleton trajectories as a composition of motion primitives. Motion programs can be automatically inferred from video and used in video editing.
sumith1896.github.io/motion2… (1/n)
Shubham Goel retweeted
Excited to share our latest work, which generates high-res images non-adversarially without the use of a discriminator. Benefits include: (1) no mode collapse and (2) no training instability.
Joint work with @PengShichong
Paper and Code: niopeng.github.io/HyperRIM/
Thread
1/n
Code for our ECCV2020 paper "Shape and Viewpoints without Keypoints" (aka UCMR) is now public at github.com/shubham-goel/ucmr!
@akanazawa
We're presenting our #ECCV2020 virtual poster today from 4-6pm (PT)! Come see us at paper #2305 to chat about learning to reconstruct 3D shapes without 3D GT!!
shubham-goel.github.io/ucmr/
Shubham Goel retweeted
If you are attending ECCV, and are lost trying to figure out what poster is when, please find a list of papers per poster session, with links to paper information pages: s-gupta.github.io/sane-eccv-…
@eccvconf
Shubham Goel retweeted
New work in #ECCV2020 learning to reconstruct a morphable shape, texture, and viewpoint from an image collection without 3D ground truth *and* 2D keypoints, allowing us to explore new categories like shoes!
Congrats to @goelshbhm!
arxiv.org/abs/2007.10982
shubham-goel.github.io/ucmr/