A Signal Analysis for Feature Points Tracking of Lip Movements

In this paper, we proposed real-time tracking of lip feature points and its signal analysis for obtaining visual information while pronouncing. By extracting and tracking the feature points of the lip, it can obtain the visual signals and information. In addition, we compared with the performance of visual and auditory signals in noisy environments. We test a performance with the four basic control commands. The recognition rate is 82.5% with the visual signals. The results showed that the visual signals have better performance than auditory with noise. We will be planning to apply a hybrid signals with both auditory and visual.