Master Thesis MSTR-2023-73

BibliographyRavi, Niranjan: Point cloud and particle data compression techniques.
University of Stuttgart, Faculty of Computer Science, Electrical Engineering, and Information Technology, Master Thesis No. 73 (2023).
76 pages, english.

The contemporary need for heightened processing speed and storage capacity has necessitated the implementation of data compression in various applications. This study encompasses a diverse array of applications, varying in scale, that need the implementation of efficient compression techniques. At present, there is no universally preferred compression technique that can outperform others across all data types. This is due to the fact that certain compression methods are more effective in compressing specific applications than others. Point cloud data finds widespread usage in diverse domains such as computer vision, robotics, and virtual as well as augmented reality. The dense nature of point cloud data presents difficulties with respect to storage, transmission, and computation. In a similar way, particle data usually contains significant amounts of particles that have been produced through simulations, experiments, or observations. The magnitude of particle data and the computational resources necessary to handle and examine such datasets can pose a formidable obstacle. To date, there has been no direct comparative analysis of compression methodologies applied to particle data and point cloud data. This study represents the initial attempt to compare these two distinct categories. The primary objective of this study is to test different compression techniques belonging to the particle and point cloud worlds and establish a standardized metric for evaluating the effectiveness of those compression methodologies. An integrated tool has been developed in this work that incorporates various compression techniques to evaluate the appropriateness of each technique for particle data and point cloud data. The assessment of compression techniques involves the consideration of particle error metrics and point cloud error metrics. Evidence from experiments in this work demonstrates that particle compressors exhibit superior performance across both tested data categories, while point cloud compressors demonstrate superior performance solely for point cloud data. Also, it reveals that the particle error metrics exhibit stringent boundaries, which are deemed necessary for the type of data they are intended to analyze. In contrast, the point cloud error metrics display more relaxed boundaries.

Full text and
other links
Department(s)University of Stuttgart, Institute of Visualisation and Interactive Systems, Visualisation and Interactive Systems
Superviser(s)Ertl, Prof. Thomas; Meese, Dr. Bernd; Reina, Dr. Guido; Gralka, Patrick
Entry dateFebruary 20, 2024
   Publ. Computer Science