Professor for Visual Computing & Artificial Intelligence at TU Munich
Co-Founder @Synthesia
From time to time, research can be quite intense. To stay sharp, I love getting some exercise - it's actually a great way to come up with new research ideas.
Also funny: quite a few PhD students joined the same gym recently, hence creating a huge incentive to not slack off :)
We released our Neural Parametric Head Models (NPHM) dataset from our #CVPR2023 paper!
It includes over 5600 high-fidelity 3D scans of human heads from 272 subjects - all publicly available!
Check it out!
https://simongiebenhain.github.io/NPHM/
We've also released the NPHM code: https://github.com/SimonGiebenhain/NPHM
We provide several examples and a small demo how to use our dataset. It's now super easy to build your own parametric face/head model!
Great work by Simon, Tobias, Markos Martin Lourdes!
Super exciting keynotes scheduled at #Eurographics2023!
Looking forward to seeing the talks by Elmar Eisemann,
Gordon Wetzstein, Ben Mildenhall, and Mirela Ben-Chen!
On my way to #Eurographics2023!
Looking forward to an exciting paper program. Also feel free to ping me if you want to catch up. I'd love to learn more about ongoing research directions :)
📢📢📢NeRSemble #SIGGRAPH'23!
We reconstruct dynamic radiance fields for high-quality novel view synthesis of human heads.
Key is a deformation field and an ensemble of multi-resolution hash encodings to model coarse & fine-scale deformations.
To train our model, we record a novel multi-view video dataset from 16 cameras containing over 4700 sequences of human heads covering a variety of facial expressions.
The dataset will be publicly available!
Video: https://youtu.be/a-OAWqBzldU
Preparing our "Introduction to Deep Learning" course!
Almost done automating myself with my #AI avatar :)
Avatar by @synthesiaIO #AI #GenerativeAI
BREAKING: #OpenAI to build a Dyson sphere around the sun to power training its next-gen #LLM on 5+ trillion A500s. Taking away sunlight until training converges seems a small price to pay :)
Rumors also have it that #GPTx will have more parameters than atoms in the universe 😉
Camera frame registration with little or no overlap?
Check out Can's ObjectMatch: We leverage object semantics to find indirect correspondences between frames. Great for any SLAM and SfM pose optimization!
Video: https://youtu.be/kuXoKVrzURk
Project: https://cangumeli.github.io/ObjectMatch/
Excited to share Norman's DiffRF: Rendering-guided 3D Radiance Field Diffusion #CVPR2023 highlight!
2D diffusion is great, but what about 3D? We show radiance field diffusion with rendering guidance for consistent and editable 3D synthesis.
Video: http://youtu.be/qETBcLu8SUk
Project: sirwyver.github.io/DiffRF
Great collaboration with Peter's group at Meta!
