Person Recognition Using Features of Still Face Images and Text-independent Speeches
Lin Huang, Hanqi Zhuang, Wenjing Zhang · 2007
An algorithm to fuse feature information of still images and text-independent speeches is proposed under a framework of Probabilistic Neural Network (PNN), and the conception of feature frame in feature space is proposed to synchronize and integrate the features from both signal sources. Experiment results on a virtual database show that the proposed method achieves a better recognition rate than either still face or speech technique involved in the test.