A Lip Reading System Based on OLSDA

Yuan-yao LU, Ying CAI, Qing-qing LIU


Lip features extraction is a very crucial part in lip reading system, accurate and effective features can promote the recognition rate of system. In this paper, we propose a features extraction method which is based on geometrical model and Optimized Locality Sensitive Discriminant Analysis (OLSDA). First, the key points in the lip area are extracted artificially, lip visual features are obtained by mathematical treatment. Then, the dimension of these features are reduced by OLSDA and processed features maintain the ability of recognize. We verify the method by SVM classifier and experimental results show that this system has a high recognition rate.


Lip reading, Geometrical model, OLSDA, SVM


Full Text:



  • There are currently no refbacks.