The Visual Computing & Artificial Intelligence Lab at TUM is looking for a student assistant (m/f/x) for a project in neural rendering of human faces starting as soon as possible.
HiWi for Creation of 3D Head Avatar Benchmark
12.12.2024, Studentische Hilfskräfte, Praktikantenstellen, Studienarbeiten
The Visual Computing & Artificial Intelligence Lab at TUM is looking for a student assistant (m/f/x) starting as soon as possible. The tasks include implementing and evaluating baselines for dynamic novel view synthesis of human heads and monocular 3D head avatar creation.
Photorealistic rendering of faces is currently a very active research field in computer vision. To further progress, we want to establish a benchmark for two common tasks: (1) photo-realistic dynamic novel view synthesis of human heads, and (2) 3D head avatar creation from monocular videos. To realize our benchmark, we are looking for a student assistant to support us flexibly 5-10h per week over a period of 3 months.
We offer:
- Insights into cutting-edge research in neural rendering for humans
- Acquiring knowledge of applied computer vision that is invaluable to many companies
- Flexible working hours (5-20h) & TUM-typical remuneration
Tasks:
- Implement and run baseline methods for dynamic novel view synthesis and monocular 3D head avatar creation
- Evaluate baselines with a shared evaluation protocol
Requirements:
- Good knowledge of PyTorch and experience with implementing neural networks
- Knowledge of 3D computer vision (3D Gaussian Splatting, camera conventions, differentiable rendering)
- High standards for code quality. Ability to write maintainable code
- Communicative and independent working person
The employment contract is initially limited to 3 months, but there is the possibility of an extension.
If you are interested, please send an email with your CV and grade reports to Tobias Kirschstein (tobias.kirschstein@tum.de).
Kontakt: tobias.kirschstein@tum.de
Mehr Information
1 |
HiWi Position Avatar Benchmark,
(Type: application/pdf,
Größe: 50.2 kB)
Datei speichern
|