Publication Details
Estimation of Gender and Age from Recorded Speech
speech processing, age estimation, gender estimation
The paper is on Estimation of Gender and Age from Recorded Speech
This document deals with gender and age estimation from recorded speech based on Gaussian Mixture Models (GMMs). Estimation process is composed of training of models representing appropriate speaker groups and following speaker classification using these trained models. Records from Czech SpeechDat(E) database are used as training and test data set. In order to reduce data size, Mel-Frequency Cepstral Coeficients (MFCC) are extracted speech. Additionally, discriminative training is applied to the trained models to provide further improvement of the results. Finally, obtained results are discussed and explained. Mistakes accompanying the estimation procedure are compared with the mistakes of subjective gender and age estimation by human listeners.
@INPROCEEDINGS{FITPUB8224, author = "Valiantsina Hubeika", title = "Estimation of Gender and Age from Recorded Speech", pages = "25--32", booktitle = "Proc. ACM Student Research competition 2006", year = 2006, location = "Prague, CZ", publisher = "Czech Technical University", ISBN = "80-01-03595-6", language = "english", url = "https://www.fit.vut.cz/research/publication/8224" }