OuluVS2: A multi-view audiovisual database for non-rigid mouth motion analysis

摘要: Visual speech constitutes a large part of our nonrigid facial motion and contains important information that allows machines to interact with human users, for instance, through automatic visual recognition (VSR) speaker verification. One the major obstacles research non-rigid mouth analysis is absence suitable databases. Those available public either lack sufficient number speakers or utterances contain constrained view points, which limits their representativeness usefulness. This paper introduces newly collected multi-view audiovisual database analysis. It includes more than 50 uttering three types importantly, thousands videos simultaneously recorded by six cameras from five different views spanned between frontal profile views. Moreover, simple VSR system has been developed tested on provide some baseline performance.