| Summary: | This paper presents works involves in developing a three-dimensional lips model to be used in a visual speech synthesis. Even though several works have been done in similar areas, but a systematic approach of identifying and determining the number of vertices used in the lips model for visual speech synthesis were not explained thoroughly. Instead of the development technique involved, this paper also present experimental design and set up to test the capabilities and flexibility of the lips in deforming speech. The work starts with the process of capturing the lips shape using 3D scanner and then the number of vertices
reduced using relevance measure technique in order to achieve the optimum number of vertices to deform all visemes in Standard Malay. Results shown in this work proves that a systematic degradation of 10,000 vertices into 60 vertices by retaining 0.0006% of the initial vertices managed to be done. We also found out that the deformation of the lips with smaller number of vertices is still remaining the similar quality of the lips deformation of the 150 or 100 vertices. Therefore, the outcome of this study proposes and justifies the appropriate number of vertices to be used in visual speech synthesis in our future work.
|
|---|