Abstract
In this paper we describe speaker and command recognition related experiments, through quantile vectors and Gaussian Mixture Modelling (GMM). Over the past several years GMM and MFCC have become two of the dominant approaches for modelling speaker and speech recognition applications. However, memory and computational costs are important drawbacks, because autonomous systems suffer processing and power consumption constraints; thus, having a good trade-off between accuracy and computational requirements is mandatory. We decided to explore another approach (quantile vectors in several tasks) and a comparison with MFCC was made. Quantile acoustic vectors are proposed for speaker verification and command recognition tasks and the results showed very good recognition efficiency. This method offered a good trade-off between computation times, characteristics vector complexity and overall achieved efficiency.
Highlights
There have been several approaches for biometry recognition using an individual’s physical characteristics, which can be acquired through signals originated from the human body or specific traits in the individual’s physiology
GonzálezI-nAtrJriAagdav ORo. bHoutgSoy,sNt,y2ls0a1F4lo, r1e1s:2a1nd| dVoizi:ca1r0ra.5-C77o2rr/a5l6L2u5is6: 1 Quantile Acoustic Vectors vs. MFCC Applied to Speaker Verification
The octile vectors were obtained from wav files in both short-time and long-time separately
Summary
There have been several approaches for biometry recognition using an individual’s physical characteristics, which can be acquired through signals originated from the human body or specific traits in the individual’s physiology. These signals could be used for several purposes, such as pathology detection, biometric identification or autonomous systems applications. In the human body context, the respiratory system is closely related with the vocal tract (which has been extensively studied) and the sub-glottal areas (a topic of detailed acoustic research as well). GonzálezI-nAtrJriAagdav ORo. bHoutgSoy,sNt,y2ls0a1F4lo, r1e1s:2a1nd| dVoizi:ca1r0ra.5-C77o2rr/a5l6L2u5is6: 1 Quantile Acoustic Vectors vs. MFCC Applied to Speaker Verification
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call