A Bioinspired Hierarchical System for Speech Recognition

摘要

Artificial speech recognition systems lack certain characteristics needed for maintaining their performance under usual conditions (background noise, continuos speedch...etc). Human auditory system has been able to solve these problems through neural evolution. In this paper a bioinspired speech recognition system, which mimic the hierarchical auditory processing is proposed. It will present the desired robustness, accuracy and spectro-temporal generalization.