Date of Award:


Document Type:


Degree Name:

Master of Science (MS)


Computer Science

Committee Chair(s)

Steve Petruzza


Steve Petruzza


Isaac Cho


John Edwards


Point clouds are widely used in various applications such as 3D modeling, geospatial analysis, robotics, and more. One of the key advantages of 3D point cloud data is that, unlike other data formats like texture, it is independent of viewing angle, surface type, and parameterization. Since each point in the point cloud is independent of the other, it makes it the most suitable source of data for tasks like object recognition, scene segmentation, and reconstruction. Point clouds are complex and verbose due to the numerous attributes they contain, many of which may not be always necessary for rendering, making retrieving and parsing a heavy task.

As Sensors are becoming more precise and popular, effectively streaming, processing, and rendering the data is also becoming more challenging. In a hierarchical continuous LOD system, the previously fetched and rendered data for a region may become unavailable when revisiting it. To address this, we use a non-persistence cache using hash-map which stores the parsed point attributes, which still has some limitations, such as the dataset needing to be refetched and reprocessed if the tab or browser is closed and reopened which can be addressed by persistence caching. On the web, popularly persistence caching involves storing data in server memory, or an intermediate caching server like Redis. This is not suitable for point cloud data where we have to store parsed and processed large point data making point cloud visualization rely only on non-persistence caching.

The thesis aims to contribute toward better performance and suitability of point cloud rendering on the web reducing the number of read requests to the remote file to access data.We achieve this with the application of client-side-based LRU Cache and Private File Open Space as a combination of both persistence and non-persistence caching of data. We use a cloud-optimized data format, which is better suited for web and streaming hierarchical data structures. Our focus is to improve rendering performance using WebGPU by reducing access time and minimizing the amount of data loaded in GPU.

Preliminary results indicate that our approach significantly improves rendering performance and reduce network request when compared to traditional caching methods using WebGPU.