SavedAdd your journal

Journals / Mehmet Akif Ersoy Üniversitesi Uygulamalı Bilimler Dergisi / 2018 / Cilt: 2 Sayı: 2

Comparison of Different Normalization Techniques on Speakers’ Gender Detection

Konuşmacı Cinsiyetinin Tespitinde Değişik Normalizasyon Tekniklerinin Kıyaslanması

Serhat İleri(Ondokuz Mayıs Üniversitesi | Ondokuz Mayıs University)0000-0002-0259-0791
Armağan Karabina(Recep Tayyip Erdoğan Üniversitesi)
Erdal Kılıç(Ondokuz Mayıs Üniversitesi | Ondokuz Mayıs University)
Pages
1–12
Publication Date
2018-09-30
🏛️Bibliographic metadata and full text access for this article are provided via DergiPark.
DergiPark Record ↗
✨

AI Summary & Key Takeaways

AI Insight
⚡ TL;DR(Too Long; Didn't Read)

As a result of study, it was observed that, the highest accuracy in speakers’ gender estimation is obtained as %98.18 from features which were normalized with Standard Deviation Normalization technique and other normalization techniques were reduced accuracy.

📌 3 Key Takeaways

🎯
Objective & Hypothesis:In this study, the effect of Short-time Mean and Variance Normalization (STMVN), Short-time Cepstral Mean and Scale Normalization (STMSN), Min-Max Normalization, Z-Score Normalization and Standard Deviation Normalization techniques on the classification performance was investigated in determining speakers’ gender.
📊
Key Findings & Data:In the study, voice records which belongs to 192 male and 192 female speakers from TIMIT data set were used as data set.
💡
Conclusion & Impact:As a result of study, it was observed that, the highest accuracy in speakers’ gender estimation is obtained as %98.18 from features which were normalized with Standard Deviation Normalization technique and other normalization techniques were reduced accuracy.
🔬Design:Konferans Bildirisi
📍Scope:Comprehensive Sample
⚙️Variables:Asgari – Azami Normalizasyonu · Z-Skor Normalizasyonu · Standart Sapma Normalizasyonu

Abstract

In this study, the effect of Short-time Mean and Variance Normalization (STMVN), Short-time Cepstral Mean and Scale Normalization (STMSN), Min-Max Normalization, Z-Score Normalization and Standard Deviation Normalization techniques on the classification performance was investigated in determining speakers’ gender. In the study, voice records which belongs to 192 male and 192 female speakers from TIMIT data set were used as data set. Features were extracted from Mel Frequency Cepstral Coefficients ( MFCC) technique by using voice records and extracted features’ dimension was reduced to Principal Component Analysis ( PCA), then normalized with different techniques. Support Vector Machine (SVM) was used as classifier. As a result of study, it was observed that, the highest accuracy in speakers’ gender estimation is obtained as %98.18 from features which were normalized with Standard Deviation Normalization technique and other normalization techniques were reduced accuracy.

Özet

Bu çalışmada Kısa-zaman Ortalama ve Değişinti Normalizasyonu (Short-time Mean and Variance Normalization - STMVN), Kısa-zaman Sepstral Ortalama ve Ölçeklendirme Normalizasyonu (Short-time Cepstral Mean and Scale Normalization - STMSN), Asgari – Azami (Min-Max) Normalizasyonu, Z-Skor (Z-Score) Normalizasyonu ve Standart Sapma (Standard Deviation) Normalizasyon tekniklerinin, konuşmacı cinsiyetinin tespitinde sınıflandırma başarımına etkisi araştırılmıştır. Çalışmada veri seti olarak TIMIT veri setindeki 192 erkek ve 192 kadın konuşmacıya ait ses kayıtları kullanılmıştır. Ses kayıtlarından Mel Frekansı Sepstral Katsayısı (Mel Frequency Cepstral Coefficient – MFCC) tekniği ile öznitelik çıkarılmış ve çıkarılan özniteliklerin boyutu Temel Bileşen Analizi (Principal component analysis – PCA) ile indirgenerek, değişik teknikler ile normalize edilmiştir. Sınıflandırıcı olarak Destek Vektör Makinesi (Support Vector Machine – SVM) kullanılmıştır. Çalışma sonucunda konuşmacı cinsiyeti tahmininde en yüksek başarımın %98.18 ile Standart Sapma Normalizasyon Tekniği ile normalize edilmiş özniteliklerden elde edildiği gözlemlenmiş olup diğer tekniklerin başarımı düşürdüğü gözlemlenmiştir.

🔖

Cite This Article

İleri, S., Karabina, A., Kılıç, E. (2018). Comparison of Different Normalization Techniques on Speakers’ Gender Detection. Mehmet Akif Ersoy Üniversitesi Uygulamalı Bilimler Dergisi, Cilt: 2 Sayı: 2, 1–12. https://doi.org/10.31200/makuubd.410625

📄 Full Text PDF Viewer

📄

PDF görüntüleyici hazırlanıyor...

References (14)

  1. Alam, M. J. vd. (2011) Comparative evaluation of feature normalization techniques for speaker verification. International Conference on Nonlinear Speech Processing. Springer Berlin Heidelberg
  2. Chen, O. T-C. & Gu, J. J. (2015) Improved gender/age recognition system using arousal-selection and feature-selection schemes. Digital Signal Processing (DSP), 2015 IEEE International Conference on. IEEE
  3. Djemili, R. vd. (2012)A speech signal based gender identification system using four classifiers. Multimedia Computing and Systems (ICMCS), 2012 International Conference on. IEEE
  4. Durukal, M. & Hocaoğlu A. K. (2015) Performance optimization on emotion recognition from speech. Signal Processing and Communications Applications Conference (SIU), 2015 23th. IEEE
  5. Heerden C. vd. (2010) Combining regression and classification methods for improving automatic speaker age recognition. Acoustics Speech and Signal Processing (ICASSP), 2010 IEEE International Conference on. IEEE
  6. Islam, M. A. (2016). GFCC-based robust gender detection. In Innovations in Science, Engineering and Technology (ICISET), International Conference on. IEEE.
  7. Khanum, S., & Firos, A. (2017). Text independent gender identification in noisy environmental conditions. In Computing, Communication and Automation (ICCCA), 2017 International Conference on. IEEE.
  8. Kizrak, M. A. & Bolat, B. (2914) Klasik Türk Müziği Makamlarının Tanınması. Akıllı Sistemlerde Yenilikler ve Uygulamaları Sempozyumu (ASYU) 2-6.
  9. Nabiyev, V. V. & Yücesoy, E. (2009) VQ Yöntemiyle Konuşmacı Cinsiyetinin Belirlenmesi. Turkish Journal of Computer and Mathematics Education Vol 1.1, 35-47.
  10. Přibil, J. vd. (2016) GMM-Based Speaker Gender and Age Classification After Voice Conversion. Sensing, Processing and Learning for Intelligent Machines (SPLINE), 2016 First International Workshop on. IEEE
  11. Yücesoy, E. & Nabiyev, V. V. (2014) Comparison of MFCC, LPCC and PLP features for the determination of a speaker's gender. Signal Processing and Communications Applications Conference (SIU), 2014 22nd. IEEE
  12. Yusnita, M. A. vd. (2017) Automatic gender recognition using linear prediction coefficients and artificial neural network on speech signal. In Control System, Computing and Engineering (ICCSCE), 2017 7th IEEE International Conference on. 2017
  13. Yücesoy, E. & Nabiyev, V. V. (2009) Gender identification of the speaker using DTW method. Signal Processing and Communications Applications Conference, SIU 2009. IEEE 17th. IEEE
  14. Yücesoy, E. & Nabiyev, V. V. (2016) Konuşmacı Yaş Ve Cinsiyetinin Gkm Süpervektörlerine Dayalı Bir Dvm Sınıflandırıcısı İle Belirlenmesi. Gazi Üniversitesi Mühendislik-Mimarlık Fakültesi Dergisi 31.3