Auditory Neural Feedback as a Basis for Representing the Temporal Cues of Speech

A fundamental component in every speech recognition system is its front-end. Most of the existing front-ends are based on variations of Fourier analysis. Another kind of front-end is based on current knowledge of the auditory periphery. A potential advantage of auditory-based front-ends is their ability to emphasize relevant (in a human-perception sense) speech information.