User voice idenfication on FPGA

User voice identification algorithms are traditionally developed on off-the-shelf digital signal processors. Recent silicon technology advances have provided the opportunity of deploying FPGAs as an alternative solution for this purpose. However, to date, the realisation of complex signal processing algorithms has been mainly based on software implementation. Inevitably, this has led to a wide gap between algorithms and hardware. This paper is aimed to explore algorithm-hardware migration technologies and, in particular, the realisation of speech feature extraction using FPGAs.