Kartik Teotia

Research

My work has focused on advancing neural representations for head avatars — from implicit volumetric models for high-quality reconstruction (HQ3DAvatar, SIGGRAPH 2024), to real-time drivable Gaussian head avatars (GaussianHeads, SIGGRAPH Asia 2024), and universal audio-driven avatars (UniGAHA, SIGGRAPH Asia 2025). More broadly, I am interested in generative models that unify photorealistic 3D facial representations with sensory signals such as video and speech.

Publications

UniGAHA: Audio-Driven Universal Gaussian Head Avatars

Kartik Teotia, Helge Rhodin, Mohit Mendiratta, Hyeongwoo Kim, Marc Habermann, Christian Theobalt

SIGGRAPH Asia 2025 Conference

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

GaussianHeads: End-to-End Learning of Drivable Gaussian Head Avatars from Coarse-to-fine Representations

Kartik Teotia, Hyeongwoo Kim, Pablo Garrido, Marc Habermann, Mohamed Elgharib, Christian Theobalt

ACM Transactions on Graphics (SIGGRAPH Asia 2024)

We propose a new method for generating dynamic, deformable human head avatars from multi-view imagery in real-time. Our coarse-to-fine facial representation enables high-fidelity expression synthesis from novel viewpoints.

paper / project page

HQ3DAvatar: High Quality Implicit 3D Head Avatar

Kartik Teotia, Mallikarjun B R, Xingang Pan, Hyeongwoo Kim, Pablo Garrido, Mohamed Elgharib, Christian Theobalt

ACM Transactions on Graphics (TOG 2024, presented at SIGGRAPH 2024)

We present a novel method for rendering photorealistic human head avatars. Our method utilizes an implicitly learned canonical space constrained using optical flow in a multiresolution hash encoding framework.

paper / 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, Christian Theobalt

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.

paper / project page

Automatic segmentation of common carotid artery in longitudinal mode ultrasound images using active oblongs

J. R. Harish Kumar, Kartik Teotia, P.K. Raj, J. Andrade, K. V. Rajagopal, C. S. Seelamantula

International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2019)

We propose a fully automated algorithm for the segmentation of common carotid artery in longitudinal mode ultrasound images using active oblongs.

paper

Education