Personal Identification with Face and Voice Features Extracted through Kinect Sensor
Eisuke Kita, Yi Zuo, Fumiya Saito, Xuanang Feng · 2016
The personal identification from the features of personal face and voice is described in this study. The face area is detected from the picture including both the face and the complicated background by using Microsoft Kinect sensor. The personal voice is also recorded from Kinect microphone array, which is used for the personal identification. The features of the personal face are calculated from the position vectors of the face parts such as eyes, nose, mouth and so on. The mel-frequency cepstrum coefficients, the logarithmic power and their related values are calculated from the personal voice. The personal identification algorithm is defined by neural network and support vector machine. The identification accuracy of the algorithms are confirmed by the face and the voice data observed from 20 examinees. The results show that the best accuracy can be observed when both face and voice data are adopted and the algorithm is defined by the neural network.