Efficient Ray Tracing For Mobile Devices

<p>The demand for mobile devices with higher graphics performance has increased substantially in the past few years. Most mobile applications that demand 3D graphics use commonly available frameworks, such as Unity and Unreal Engine. In mobile devices, these frameworks are built on top of OpenGL ES and use a graphics technique called rasterization, a simple concept that yields good performance without sacrificing graphic quality. However, rasterization cannot easily handle some physical phenomena of light (i.e. reflection and refraction). In order to support such effects, the graphics framework has to emulate them, thereby leading to suboptimal results in term of quality.</p> <p>Other techniques, such as ray tracing, do not require such emulation to be implemented, as the aforementioned phenomena of light are inherently considered. In this work, we first design and implement a 3D renderer that uses ray tracing to generate high quality graphics for mobile devices. Our implementation yielded high quality results but at a high computational cost, which impacted performance. To alleviate this problem, we then developed special algorithms and data structures to substantially improve the performance of the rendering engine. In our experiments, we achieved frame rates that were 7 to 15 times faster than the brute force approach. Being able to render high quality graphics with good performance can potentially revolutionize the mobile gaming industry. To the best of our knowledge, this has never before been implemented in commonly available devices, such as smartphones and tablets.</p>

Name thesis:degree_name

Master of Science in Computer Science (MS)

Level thesis:degree_level

Thesis

Year dc:date.available

2014

Author dc:creator

• De Melo Aroxa, Rafael

Advisor dc:contributor.advisor

• Gauch, John M.

Contributors dc:contributor

• Thompson, Craig W.
• Yan, Tingxin

• Applied sciences
• Computer graphics
• Mobile
• Raytracing
• Graphics and Human Computer Interfaces

Repository record dc:identifier

https://scholarworks.uark.edu/etd/1055

OAI identifier oai:identifier

oai:scholarworks.uark.edu:etd-2054