Summer Semester 2027 – fixed term of 3–6 months
YOUR TASKS:
-
You review and analyze state-of-the-art methods for generating LiDAR data from camera images and videos, including approaches based on deep learning, diffusion models, neural rendering and depth estimation.
-
You develop and evaluate a pipeline for generating LiDAR-like point clouds from synchronized multi-camera systems.
-
You adapt and optimize selected methods to reproduce the characteristics of a SICK LiDAR sensor.
-
You investigate the potential and limitations of existing RGB-to-LiDAR approaches for industrial applications.
-
You evaluate the generated data in terms of geometric accuracy, robustness and suitability for downstream AI applications.
-
You document and present your findings and recommendations.
The exact scope of the project will be tailored to your background and the requirements of your internship, bachelor's thesis or master's thesis.
YOUR PROFILE:
-
Studies in computer science, robotics, electrical engineering, mathematics or a related field
-
Interest in computer vision, 3D perception and sensor fusion
-
Good programming skills in Python
-
Basic knowledge of machine learning and deep learning
-
Experience with PyTorch or a similar framework is an advantage
-
Ability to work independently and analyze scientific literature
Contact: Sarah Disch