Files in this item



application/pdfELLIS-DISSERTATION-2018.pdf (7MB)Restricted Access
(no description provided)PDF


Title:A high performance vector rendering pipeline
Author(s):Ellis, Apollo Isaac Orion
Director of Research:Hart, John C.
Doctoral Committee Chair(s):Hart, John C.
Doctoral Committee Member(s):LaValle, Steven M.; Patel, Sanjay J.; Hunt, Warren A
Department / Program:Computer Science
Discipline:Computer Science
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):Vector Graphics, Rendering, GPU, Cloud Gaming, Anti-aliasing, VR, Real-time
Abstract:Vector images are images which encode visible surfaces of a 3D scene, in a resolution independent format. Prior to this work generation of such an image was not real time. As such the benefits of using them in the graphics pipeline were not fully expressed. In this thesis we propose methods for addressing the following questions. How can we introduce vector images into the graphics pipeline, namingly, how can we produce them in real time. How can we take advantage of resolution independence, and how can we render vector images to a pixel display as efficiently as possible and with the highest quality. There are three main contributions of this work. We have designed a real time vector rendering system. That is, we present a GPU accelerated pipeline which takes as an input a scene with 3D geometry, and outputs a vector image. We call this system SVGPU: Scalable Vector Graphics on the GPU. As mentioned vector images are resolution independent. We have designed a cloud pipeline for streaming vector images. That is, we present system design and optimizations for streaming vector images across interconnection networks, which reduces the bandwidth required for transporting real time 3D content from server to client. Lastly, in this thesis we introduce another added benefit of vector images. We have created a method for rendering them with the highest possible quality. That is, we have designed a new set of operations on vector images, which allows us to anti-alias them during rendering to a canonical 2D image. Our contributions provide the system design, optimizations, and algorithms required to bring vector image utilization and benefits much closer to the real time graphics pipeline. Together they form an end to end pipeline to this purpose, i.e. "A High Performance Vector Rendering Pipeline."
Issue Date:2018-12-05
Rights Information:Copyright 2018 Apollo Ellis
Date Available in IDEALS:2019-02-08
Date Deposited:2018-12

This item appears in the following Collection(s)

Item Statistics