3D Point Cloud Compression / Microsoft Kinect Video Compression
3D data is becoming one of the most widely used data structures for representing scenes, enabling a six degrees of freedom (6DoF) viewing experience. This type of data has found applications across various fields, including immersive media, autonomous driving, and healthcare. We are offering two Master’s theses related to 3D data:
1. In collaboration with MaD Lab, this Master’s thesis focuses on exploring compression techniques for Kinect’s depth video. The project investigates the impact of compression on the performance of a depression detection model by comparing results from compressed and uncompressed Kinect videos.
2. Point clouds, which typically consist of millions of 3D points, require substantial storage capacity. Therefore, the development of efficient Point Cloud Compression (PCC) techniques is essential for the practical deployment of point clouds in various applications. This thesis examines the integration of point cloud upsampling and downsampling within the compression process, with an emphasis on improving quality, particularly in low-bitrate scenarios. (Co-supervised by M.Sc. Marina Ritthaler, Room 06.019, marina.ritthaler@fau.de)
Professor:
Prof. Dr.-Ing. André Kaup
Supervisior:
M.Sc. Dat Nguyen, Room 06.026, dat.thanh.nguyen@fau.de
Prerequisites:
Python, Deep Learning,
Pytorch and Pytorch Lightning (optional)
Available:
Immediately