Proceedings of the 30th Annual International Computer Software and Applications Conference (COMPSAC '06),
IEEE Computer Society Press, Los Alamitos, CA, pp. 429-438 (2006)

Reference Models and Automatic Oracles for the Testing of
Mesh Simplification Software for Graphics Rendering
1

W.K. Chan 2 , S.C. Cheung 2 , Jeffrey C.F. Ho 3 , and T.H. Tse 4

[paper from IEEE Xplore | paper from IEEE digital library | technical report TR-2006-04]

 ABSTRACT

Software with graphics rendering is an important class of applications. Many of them use polygonal models to represent the graphics. Mesh simplification is a vital technique to vary the levels of object details and, hence, improve the overall performance of the rendering process. It progressively enhances the effectiveness of rendering from initial reference systems. As such, the quality of its implementation affects that of the associated graphics rendering application. Testing of mesh simplification is essential towards assuring the quality of the applications. Is it feasible to use the reference systems to serve as automated test oracles for mesh simplification programs? If so, how well are they useful for this purpose?

We present a novel approach in this paper. We propose to use pattern classification techniques to address the above problem. We generate training samples from the reference system to test samples from the implementation. Our experimentation shows that the approach is promising.

Keywords: Test oracles, software testing, mesh simplification, graphics rendering, pattern classification reference models.

1. This research is supported in part by a grant of the Research Grants Council of Hong Kong (project no. 714504) and a grant of The University of Hong Kong.
2. Department of Computer Science, Hong Kong University of Science and Technology, Clear Water Bay, Hong Kong.
3. Department of Computer Science, The University of Hong Kong, Pokfulam, Hong Kong.
4. (Corresponding author.)
Department of Computer Science, The University of Hong Kong, Pokfulam, Hong Kong.
Email:

 EVERY VISITOR COUNTS:

  Cumulative visitor count