Benchmarking Feature Selection Techniques on the Speaker Verification Task - PowerPoint PPT Presentation

1 / 15
About This Presentation
Title:

Benchmarking Feature Selection Techniques on the Speaker Verification Task

Description:

Speaker verification is a typical detection task: ... classifier to score subsets of features according to their discriminative power ... – PowerPoint PPT presentation

Number of Views:127
Avg rating:3.0/5.0
Slides: 16
Provided by: tgan
Category:

less

Transcript and Presenter's Notes

Title: Benchmarking Feature Selection Techniques on the Speaker Verification Task


1
Benchmarking Feature Selection Techniques on the
Speaker Verification Task
  • T. Ganchev, P. Zervas, N. Fakotakis, G.
    Kokkinakis
  • Wire Communications Laboratory,
  • University of Patras, Greece

2
Presentation Outline
  • Problem definition
  • Approaches for feature evaluation
  • Feature ranking techniques
  • Feature selection
  • Experimentations
  • Conclusion

3
Problem Definition
  • Speaker verification is a typical detection task
  • Does the speaker possess the identity s/he
    claims?
  • Our problem
  • The speech features contribute differently for
    speaker differentiation (i.e. have different
    discrimination power)
  • Our target
  • To identify and keep all potentially relevant
    features

4
Approaches for Feature Evaluation
  • Filters act on pre-processing step
  • Wrappers utilize the classifier to score
    subsets of features according to their
    discriminative power
  • Embedded methods perform feature selection
    during the training process

5
Feature ranking techniques 1/4
  • Information Gain (IG) attribute evaluation,
  • Gain Ratio (GR) attribute evaluation,
  • Symmetrical Uncertainty (SU) attribute
    evaluation,
  • Correlation-based Feature Selection (CFS),
  • Support Vector Machine Recursive Feature
    Elimination (SVM-RFE).

6
Feature ranking techniques 2/4
The Entropy of Y is defined as
The Entropy of Y after observing X is
7
Feature ranking techniques 3/4
  • Information Gain (IG) ? mutual information
  • Gain Ratio (GR)
  • Symmetrical Uncertainty (SU)

8
Feature ranking techniques 4/4
  • Correlation-based Feature Selection
  • SVM Recursive Feature Elimination

train the SVM classifier by optimizing the
weights w.r. the cost function
compute the ranking criterion for all features
remove the feature with smallest ranking
criterion
9
Feature selection
  • IG, GR, and SU does not perform feature selection
    but only feature ranking,
  • we experimented with the forward selection and
    bi-directional search.
  • In the case of CFS, we used genetic search of the
    feature space.
  • Finally, we performed full search to find out the
    appropriate number of attributes.

10
Experimental Setup
  • The WCL-1 speaker verification system
  • PNN-based expert for each target user
  • Speech parameterization
  • logarithm of fundamental frequency,
  • logarithm of energy,
  • 31 Mel-Frequency Cepstral Coefficients (where the
    0-th coefficient is excluded)
  • The 2001 NIST SRE database
  • 74 male speakers, 2 minutes of training data
  • various transmission channels TDMA, CDMA,
    Cell, GSM, Land
  • various length of trials 00-15, 16-25,
    26-35, 36-45, and 46-60 sec
  • different environmental conditions inside,
    outside, vehicle

11
Experimental results Feature ranking
  • Feature ranking
  • IG - Information Gain,
  • GR - Gain Ratio,
  • SU - Symmetrical Uncertainty,
  • CFS - Correlation-based Feature Selection,
  • SVM-RFE - Support Vector Machine Recursive
    Feature Elimination,
  • Ref - Genuine feature set.

12
Experimental results All experiments
13
Experimental results SV performance
  • Comparison among the most successful feature sets
    for the evaluated feature ranking methods.

14
Conclusion
  • We evaluated the relevance of a set of speech
    features for the task of speaker verification.
  • Feature subsets obtained from various ranking
    techniques were compared
  • Useful knowledge was gained regarding features
    that do not contribute significantly to the task.
  • Several subsets were observed to provide a
    competitive performance, which opens ways to
    lower the computational and memory demands

15
Comments and Questions
Thank you for your attention ! Any Comments or
Questions will be welcomed...
Write a Comment
User Comments (0)
About PowerShow.com