Niladri Dutt (@niladridutt) header
Niladri Dutt (@niladridutt) profile picture

Niladri Dutt

@niladridutt

ML PhD @ucl | @adobe | ex @nvidia | @ELLISforEurope | Interested in visual generation

 London, UK
niladridutt.com
Joined July 2012
  • Tweets165
  • Following480
  • Followers350
28Media
Photo by Niladri Dutt (@niladridutt)Photo by Niladri Dutt (@niladridutt)Photo by Niladri Dutt (@niladridutt)Photo by Niladri Dutt (@niladridutt)Photo by Niladri Dutt (@niladridutt)Photo by Niladri Dutt (@niladridutt)Photo by Niladri Dutt (@niladridutt)Photo by Niladri Dutt (@niladridutt)Photo by Niladri Dutt (@niladridutt)Photo by Niladri Dutt (@niladridutt)
🎉 LoST: Level of Semantics Tokenization (CVPR 2026) What if 3D generation didn’t start from coarse geometry, but from semantics? We introduce LoST, a semantic-first 3D tokenizer that orders tokens by semantic salience, so even very short prefixes can already decode into complete, plausible, recognizable 3D shapes 🧩 Early tokens capture the principal semantics. Later tokens refine the details. Result: Generate 3D shapes starting from as few as 1 token! The goal is to have high fidelity shapes even with very few tokens, with more tokens semantic alignment increases. Project: lost3d.github.io/ Paper: arxiv.org/abs/2603.17995 @CVPR @AdobeResearch @Adobe
Image from Niladri Dutt (@niladridutt)
4
20
117
6,659
Really cool work and shows how one can enable various levels of detail during generation. Checkout our #CVPR 2026 work on 3D generation-- Level of Semantics Tokenization for 3D Shapes lost3d.github.io/ Code is now released.
Gordon Wetzstein (@GordonWetzstein) profile pictureGordon Wetzstein
@GordonWetzsteinOct 6
Diffusion models spend the same compute on a blank wall as on a face. But you often know in advance where the detail will be. Introducing Level-of-Token (LoT) Diffusion: we turn that knowledge into a multiresolution token layout, with fine tokens where detail is needed and coarse tokens elsewhere. 1/9🧵
4
222
Excited to announce the @TheBMVA Symposium on World Models! 🌍 📍 London | 🗓 Nov 18, 2026 We have an amazing lineup of speakers joining us! Check out the full program, registration, and poster submission details:bmva.org/meetings/26-11-18-W… #AI #worldmodels #computervision
Image from Niladri Dutt (@niladridutt)
1
14
1,474
Join us at the Visual Concepts Workshop at CVPR!
Kenny Jones (@RKennyJones) profile pictureKenny Jones@RKennyJonesJun 2
The Visual Concepts Workshop is coming soon to @CVPR 2026! Join us on June 4 for an exciting discussion on all aspects of visual concepts: from discovery and reasoning, to controllable generation, robotics, and more! Hope to see you in Denver! #CVPR
Image from Kenny Jones (@RKennyJones)
3
246
#GenAI #3DGeneration #CVPR #tokenization #3D
1
319
A really refreshing take on 3D generation: instead of building shapes from coarse geometry, build them from semantics. huggingface.co/papers/2603.1…
Image from Daniel Cohen-Or (@DanielCohenOr1)
3
27
1
133
7,015
Excited to share the 3rd Workshop on Visual Concepts at @CVPR 2026! Call for papers is now open till April 1. We welcome submissions on the following topics. See our website for more info: sites.google.com/stanford.ed… Join us in Denver! #CVPR2026
Image from Niladri Dutt (@niladridutt)
6
23
4,003
🎉Thrilled to share that CABS is accepted to #CVPR2026! See you in Denver! Everything is now open-sourced: Webpage:cabs-vlp.github.io/ DataConcept:huggingface.co/datasets/beth… Paper:arxiv.org/abs/2511.20643 Code:github.com/bethgelab/cabs Excited to see what the community builds!
Adhiraj Ghosh (@adhiraj_ghosh98) profile pictureAdhiraj Ghosh@adhiraj_ghosh9826 Nov 2025
🚨Current data curation results in the creation of static datasets and the use of model-based filters that induce many biases. Can we fix this? We propose ✨CABS✨, a flexible concept-aware online batch curation method that improves CLIP pretraining! arxiv.org/abs/2511.20643 🧵👇
Image from Adhiraj Ghosh (@adhiraj_ghosh98)
6
2
26
9,822
Niladri Dutt retweeted
Joy Hsu (@joycjhsu) profile picture
Joy Hsu
@joycjhsuJan 21
Happy to bring back the 3rd Workshop on Visual Concepts at @CVPR 2026! Call for papers is now open. We welcome submissions on the following topics. See our website for more info: sites.google.com/stanford.ed… Join us in Denver!
Image from Joy Hsu (@joycjhsu)
1
16
1
95
16,262
Niladri Dutt retweeted
Remy Sabathier (@RemySabathier) profile picture
Remy Sabathier
@RemySabathierJan 22
Excited to introduce 🎬ActionMesh, a fast model transforming any video → high-quality animated 3D mesh ! Generate animated mesh seamlessly importable into any 3D software in less than a minute. 🤗Try it out: huggingface.co/spaces/facebo… 🌐Project Page: remysabathier.github.io/acti… 📄Paper: remysabathier.github.io/acti… 💻Code: github.com/facebookresearch/… #Video4D #GenAI #3DGeneration @Meta @RealityLabs @AIatMeta @ucl
7
82
6
592
42,964
Animation brings characters to life 🎬✨ Introducing our #SIGGRAPHAsia '25 (TOG) paper: Self-Supervised Motion Fields (SMF) that can transfer motion from any source (even 2D video) to any stylized character! 🚫 No Rigging 🚫 No Templates (SMPL) ✅ Any source 🧵Thread👇
2
1
7
486
more replies
It even works on multi-legged animals and can handle very long motion without drift
1
82
Project website: motionfields.github.io/ Paper: arxiv.org/abs/2504.04831 Code: github.com/sanjeevmk/SMF
66
Can an LLM learn professional photo retouching by solving puzzles? Our new work, MonetGPT, presented at #SIGGRAPH2025, shows it can. By solving visual puzzles, our MLLM becomes operation-aware and develops image aesthetics. Check out our blog from @AdobeResearch & code below👇 #AI #MLLM
2
7
321
✍️ Blog Post: research.adobe.com/news/adob… 💻 GitHub: github.com/niladridutt/monet… 🌐 Project Website: monetgpt.github.io/
1
1
134
Check out this thread to learn more teapaw.org/niladridutt/status/192…
Niladri Dutt (@niladridutt) profile pictureNiladri Dutt@niladridutt27 May 2025
🧵1/10 Excited to share our @siggraph paper "MonetGPT: Solving Puzzles Enhances MLLMs' Image Retouching Skills" 🌟 We explore how to make MLLMs operation-aware by solving visual puzzles and propose a procedural framework for image retouching #SIGGRAPH #MLLM
Image from Niladri Dutt (@niladridutt)
2
142
🧵1/10 Excited to share our @siggraph paper "MonetGPT: Solving Puzzles Enhances MLLMs' Image Retouching Skills" 🌟 We explore how to make MLLMs operation-aware by solving visual puzzles and propose a procedural framework for image retouching #SIGGRAPH #MLLM
Image from Niladri Dutt (@niladridutt)
1
2
1
9
675
more replies
🧵9/10 We quantitaively evaluate on the Adobe5k dataset as well as conduct user studies by expert and novice users. Our evaluations show that MonetGPT outperforms open-source alternatives and performs comparably to Google Photos AutoEnhance (closed-source).
Image from Niladri Dutt (@niladridutt)
1
131
🧵10/10 Lastly, huge thanks to my co-advisors Niloy and Duygu! For more details check out our paper below- 🌐 Project Website: monetgpt.github.io/ 📄 Arxiv: arxiv.org/abs/2505.06176
143
Load more