Learning vector quantization in text-independent automatic speaker recognition
T.E.F. Filho, Ronaldo Messina, Euvaldo F. Cabral · 2002
In this paper is reported a comparison among the learning vector quantization (LVQ) and two other common approaches to text-independent speaker recognition, namely Gaussian mixture models (GMM) and vector quantization (VQ). The LVQ method uses neural nets. The results shows that it is less efficient in terms of recognition scores than the GMM.