Generating High-Resolution Fashion Model Images Wearing Custom Outfits pdf: arxiv.org/pdf/1908.08847… abs: arxiv.org/abs/1908.08847

Day 13 of our modified StyleGAN training - things are getting good. The fashion models are much more canny, flicker is reduced. Occasional merging of two white shoes into one - it is by now "not a bug but a feature" 😎 Pose, face, garment set-conditioned Zalando GAN, 1024 px.

Personal news: I'll be joining UC San Diego Halıcıoğlu Data Science Institute as an assist. professor in Fall 2021. Looking fwd to joining the amazing colleagues. Huge thanks to mentors/collaborators Eric Xing, Russ Salakhutdinov, Tom M Mitchell, @dannydanr, Andrew Gordon Wilson, Le Song etc for the invaluable support!!

Exciting news! We launched an iPhone app for 3D human scanning with a very simple UX. Just turn around 360. You can export yourself to GTA V and Second Life or buy custom-fit jeans. Link: apple.co/2FcBZ7B

Cool!

Our SIGGRAPH Asia 2020 paper is out! Our method is the first to track and reconstruct the geometry of two interacting hands from monocular RGB video in real-time. handtracker.mpi-inf.mpg.de/projects/RGB2H…

very cool

Generating dressed photo models: visualize with a conditional GAN combinatorially many fashion outfit combinations. Join us for our Expo Demonstration: GAN Applications in Fashion Article Design and Outfit Rendering, on Monday Dec. 7th, 9AM CET. #NeurIPS2020 Zalando Research

Checkout Xinyue Wei 's latest work, which presents a transformer-based model that reconstructs a high-fidelity 3D mesh from 4 (sparse) input images in less than one second! arxiv.org/abs/2404.12385

#ECCV2024 We propose Generalizable Human Gaussians (GHG), a method that can synthesize high-quality novel view renderings of arbitrary human subject from sparse multi-view images. Website: humansensinglab.github.io/Generalizable-… Paper: arxiv.org/abs/2407.12777

I'm pretty sure everyone wants a 3D Gaussian avatar of Ronaldo or maybe Donald Trump from a single image, right? Don't despair - the new avatar paper "Human-VDM" is here to help you out! Human-VDM: Learning Single-Image 3D Human Gaussian Splatting from Video Diffusion Models.

Cool

Paper: arxiv.org/abs/2407.12777 Project: humansensinglab.github.io/Generalizable-… Code: github.com/humansensingla… Youtube: youtu.be/PyST3-dfmfI?si…

#SIGGRAPHAsia2024 We train a generative model with synthetic textures to rectify distortions in the input clothing. This yields a flat texture map with PBR materials, allowing for realistic relighting of the garment under various lightings. Project: shorturl.at/f77un

accepted at NeurIPS Conference, lets meet up in Vancouver 😁

🗣️Call for papers🗣️ ✨the 2nd GenAI4Health Workshop NeurIPS Conference 🌐genai4health.github.io Join us at NeurIPS2025 for the coolest intersection between generative AI (today’s most exciting technology) and human health (our most vital priority)! 👉🧵

#ICCV2025 Introducing💡LightSwitch💡- A multi-view material-relighting diffusion pipeline that directly and efficiently relights any number of input images to a target lighting & do 3D asset relighting with gaussian splatting! 🧵

SIGGRAPH is over, but the papers remain! If you missed it, our team presented Echoes of the Coliseum: Towards 3D Live Streaming of Sports Events Check it out here: humansensinglab.github.io/basket-multivi… doi.org/10.1145/3731214

We have extended the deadline for submissions GenAI4Health to August 31! Submit your papers and lets meet up at NeurIPS Conference 😊

We are hiring for Infra + Research Engineer roles! 🚀 Apply through our site or reach out directly if you're interested.