作者: N. Eveno , A. Caplier , P.-Y. Coulon
DOI: 10.1109/ICARCV.2002.1234982
关键词:
摘要: Lip segmentation is an essential stage in many multimedia systems such as videoconferencing, lip reading, or low bit rate coding communication systems. In this paper, we propose accurate and robust algorithm. First, the mouth region several characteristic points are detected by using "hybrid edges" (which combine colour intensity information) a priori knowledge about morphology. Corners position, which crucial, provided coarse-to-fine process. Then, parametric model fitted on lips. We consider that boundary composed of independent cubic polynomial curves. It gives complexity global flexible enough to reproduce specificity very different shapes. Compared existing models, it brings significant accuracy realism improvement. Moreover, ensures convergence towards edges because parts independent.