Felix Wimbauer

@felixwimbauer

Stealth AI for Bio | PhD @tumcvg & @UniofOxford, Prev Intern @Google & @Meta GenAI, fwmb.github.io

München, Bayern

Joined November 2020

Replies Reposts Media

Pinned
Felix Wimbauer
@felixwimbauer
May 13, 2025
Can you train a model for pose estimation directly on casual videos without supervision? Turns out you can! In our #CVPR2025 paper AnyCam, we directly train on YouTube videos and achieve SOTA results by using an uncertainty-based flow loss and monocular priors! ⬇️
00:00
4
Felix Wimbauer
@felixwimbauer
Apr 23, 2025
Check out our recent #CVPR2025 paper AnyCam, a method for pose estimation in casual videos! 1️⃣ Can be directly trained on casual videos without the need for 3D annotation. 2️⃣ Based around a feed-forward transformer and light-weight refinement. ♦️ fwmb.github.io/anycam/
00:00
5
Felix Wimbauer
@felixwimbauer
Jun 11, 2024
Diffusion models are often prohibitively slow. 😢 In our #CVPR2024 paper "Accelerating Diffusion Models through Block Caching", we present a technique to make SOTA models like LDM up to 1.8x faster without sacrificing quality.🚀 Check out our video here: youtu.be/pQjJmDltIDk
00:00
3
Felix Wimbauer
@felixwimbauer
Mar 18, 2024
Check out this new exciting paper by our lab! We will present it at @3DVconf in Davos this week. Check out our 🗣️ Spotlight presentation and Poster during Session 4, Wednesday March 20th 3:30pm.
ahayler
@adrihayler
Mar 18, 2024
🚨3DV 2024 Spotlight🚨 Turns out you don't need 3D supervision for Semantic Scene Completion!🙅‍♂️ Our new method S4C achieves competitive performance using Neural Fields and 2D Reconstruction Losses during training. 📈 🗣️ Davos, Wed, March 20, 3:30pm 🌐 ahayler.github.io/publications/s…
00:00
Felix Wimbauer
@felixwimbauer
Feb 8, 2024
Hi @3DVconf , we are already excited for attending the conference. We have uploaded our paper alongside the copyright form to the collection website. However, we just received an email stating that it hadn't been received yet. Is there anything we have to do still? Best,