Speech Recovery Based On Auditory Radar and Webcam

This paper presents a speech recovery technology based on a 24-GHz portable auditory radar and webcam for noncontact robust speech recognition, recovery and surveillance. The time-varying vocal vibration signal obtained by the continuous-wave auditory radar is used as the sound source excitation while the fitted formant frequency extracted by webcam is used as the vocal tract resonance characteristics to synthesize and recover speech. Experiments of reading single English character are carried out. Compared with microphone-recorded results, the speech recovery technology can accurately extract the formant frequency and recover speech effectively. Subject evaluation results show a high relatively consistency between the synthesized speech and original acoustic speech.