Robust Cochlear-Model-Based Speech Recognition

Accurate speech recognition can provide a natural interface for human–computer interaction. Recognition rates of the modern speech recognition systems are highly dependent on background noise levels and a choice of acoustic feature extraction method can have a significant impact on system performance. This paper presents a robust speech recognition system based on a front-end motivated by human cochlear processing of audio signals. In the...