This research project focuses on a distributed visualization pipeline which combines VR technologies and remote computing resources through high speed networks, so that large-scale scientific datasets can be visualized in real-time on local VR devices which provide immersive comprehension to scientific visualization users.

The pipeline represents a sort-middle computing architecture where 3D point samples are introduced as an intermediate format of data which flow through the pipeline. The straightforward functionality decomposition of point-based graphics enables flexible and balanced workload distribution through the computation pipeline. Each subsystem of the distributed system can perform scalable computing. Scalability makes the system adaptive to the available computing and visualizing resource configurations.

At the implementation level, a variety of point-based sampling, packing and rendering algorithms with different levels of computational complexity are studied through this thesis. The pipeline subsystem computing, communication, and coupling schemes and their supporting algorithms are discussed. As customized instances of the proposed visualization framework, case studies to visualize various datasets with different characteristics, such as triangle meshes, mathematical models and volumes, are examined to show improved VR interaction for large-scale VR problems.