Michal Geyer
I'm a graduate student in computer vision at the Weizmann Institute of Science, under the supervision of Dr. Tali Dekel.
My research is focused on generative models, their internal representations, and how they can be used to enhance our visual reality. I'm particularly interested in video editing and synthesis.
I'm also interested in computational linguistics and in multimodal learning (vision & language).
Before getting into computer science I was a professional contemporary dancer. See some selected pieces I performed below.
Publications
TokenFlow: Consistent diffusion features for consistent video editing
ICLR 2024 | project page
Michal Geyer*, Omer Bar-Tal*, Shai Bagon, Tali Dekel
*stands for equal contribution
Plug-and-Play diffusion features for image-to-image translation
CVPR 2023 | project page
Narek Tumanyan*, Michal Geyer*, Shai Bagon, Tali Dekel
*stands for equal contribution
Neural Congealing: Aligning images to a joint semantic atlas
CVPR 2023 | project page
Dolev Ofri-Amar, Michal Geyer, Yoni Kasten, Tali Dekel
Minimum description length recurrent neural networks
TACL 2022 | paper
Nur Lan, Michal Geyer, Emmanuel Chemla, Roni Katzir
Dance
GIRLS by Roy Assaf
Performed in Trefo Theater (Budapest, Hungary) and in Suzan Dallal center
(Tel Aviv, Israel )
Colonia by Talia Beck
Toured in Israel as part of curtain up dance festival
The Ballet Junior de Genève is a training aimed at the development of a young dancer destined to a professional career with his own artistic identity.