ID 2523: Multi-camera Pose estimation

Research Project

We are offering two Master’s projects (10 ECTs) focused on the practical application of 3D human pose estimation.  This project connects hardware setup with software applications by creating a multi-camera data collection system and deploying advanced pose-estimation models to analyze the captured data.

The student will be responsible for the end-to-end process of creating a custom dataset and running baseline algorithms on it. Key tasks include:

• System Setup: Designing and physically configuring a multi-camera recording environment.
• Calibration & Synchronization: Applying multi-camera calibration techniques to accurately map 3D spatial geometry and ensuring precise video synchronization.
• Model Implementation: Setting up, running, and evaluating several existing 3D human pose estimation models (e.g., SMPL-based methods or Vision Transformers) on the newly collected multi-view data.

Candidate Profile: We do not expect applicants to have complete, pre-existing expertise in 3D geometry or advanced deep learning frameworks. We are primarily looking for students with a genuine interest in the domain, a proactive approach to learning, and the motivation to see a project through from the physical setup to the software execution.

Ideal candidates will possess:

• Experience or interest in computer vision, motion capture, and practical 3D geometry.

• A willingness to engage in hands-on work with physical hardware (cameras, tripods, lighting) in addition to software tasks.

• Programming proficiency (e.g., Python) to assist with scripting data processing pipelines and running existing codebases.

• A structured approach to problem-solving and an eagerness to troubleshoot technical challenges.

Supervisors

Amirreza Asemanrafat, M. Sc.

Researcher & PhD Candidate

Please send me an email (amirreza.asemanrafat@fau.de) with your resume and transcripts to apply for the topic. We will then get in contact with you if we are interested.