MPI-INF Logo
Homepage

Contact

Mohit Mendiratta

Mohit Mendiratta

Max-Planck-Institut für Informatik
Department 6: Visual Computing and Artificial Intelligence
 office: Campus E1 4, Room 221
Saarland Informatics Campus
66123 Saarbrücken
Germany
 email: [email protected]
 phone: +49 681 9325-4532

Publications

GRMM: Real-Time High-Fidelity Gaussian Morphable Head Model with Learned Residuals
Mohit Mendiratta, Mayur Deshmukh, Kartik Teotia, Vladislav Golyanik, Adam Kortylewski, Christian Theobalt

International Conference on 3D Vision 2026 (3DV)

Our method, GRMM (Gaussian Residual Morphable Model), generates high-fidelity 3D head avatars in real time by combining a mesh-based morphable prior with learned Gaussian residuals. This design captures fine geometric and appearance details while maintaining disentangled control over identity and expression, enabling photorealistic and expressive facial synthesis.
[paper] [project page]
UniGAHA: Audio-Driven Universal Gaussian Head Avatars
Kartik Teotia, Helge Rhodin, Mohit Mendiratta, H. Kim, Marc Habermann, Christian Theobalt

ACM SIGGRAPH Asia Conference Proceedings (SIGGRAPH Asia 2025)

Our method synthesizes photorealistic 3D head avatars directly from speech. The Universal Head Avatar Prior (UHAP) adapts to new and unseen users from minimal input data, ensuring accurate lip synchronization and natural upper facial motion while preserving speaker identity.
[paper] [project page]
Zero-Shot Video Semantic Segmentation based on Pre-Trained Diffusion Models
Qian Wang, Abdelrahman Eldesokey, Mohit Mendiratta, Fangneng Zhan, Adam Kortylewski, Christian Theobalt, Peter Wonka

CVPR 2025

We introduce the first zero-shot approach for Video Semantic Segmentation (VSS) based on pre-trained diffusion models.
[paper] [project page]
TEDRA: Text-based Editing of Dynamic and Photoreal Actors
Basavaraj Sunagad, Heming Zhu*, Mohit Mendiratta*, Adam Kortylewski, Christian Theobalt, Marc Habermann (* equal contribution)

3DV 2025

Over the past years, significant progress has been made in creating photorealistic and drivable 3D avatars solely from videos of real humans. However, a core remaining challenge is the fine-grained and user-friendly editing of clothing styles by means of textual descriptions. To this end, we present TEDRA, the first method allowing text-based edits of an avatar, which maintains the avatar's high fidelity, space-time coherency, as well as dynamics, and enables skeletal pose and view control.
[paper] [video] [project page]


AvatarStudio: Text-driven Editing of 3D Dynamic Human Head Avatars
Mohit Mendiratta, Xingang Pan*, Mohamed Elgharib*, Kartik Teotia, Mallikarjun B R, Ayush Tewari, Vladislav Golyanik, Adam Kortylewski and Christian Theobalt (* equal contribution)

ACM Transactions on Graphics (SIGGRAPH Asia 2023)

AvatarStudio uses text-based editing on dynamic head avatars, employing NeRF and a text-to-image diffusion model with multiple keyframes, achieving superior personalized, 3D- and time-consistent edits in visual and numerical assessments.
[paper] [project page]
Egocentric Videoconferencing
M. Elgharib*, M. Mendiratta*, J. Thies, M. Nießner, H-P. Seidel, A. Tewari, V. Golyanik and C. Theobalt (* equal contribution)

ACM Transactions on Graphics (SIGGRAPH Asia 2020)

We introduce a method for egocentric videoconferencing that enables handsfree video calls, for instance by people wearing smart glasses or other mixedreality devices.
[paper] [video] [project page]


Recent Positions

Education