Nonlinear kernel nuisance attribute projection for speaker verification

Xianyu Zhao, Yuan Dong, Hao Yang, Jian Zhao, Liang Lu, Haila Wang

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract / Description of output

Nuisance attribute projection (NAP) was successfully applied in SVM-based speaker verification systems to improve performance by doing projection to remove dimensions from the SVM feature space that cause unwanted variability in the kernel. Previous studies of NAP were focused mainly on linear and generalized linear kernel SVMs. In this paper, NAP in nonlinear kernel SVMs, e.g. polynomial or Gaussian kernels, are investigated. Instead of doing explicit feature expansion and projection in high-dimension feature space, kernel principal component analysis is employed to find nuisance dimensions; and, NAP is carried out implicitly by incorporating it into some compensated kernel functions. Experimental results on the 2006 NIST SRE corpus indicate the effectiveness of such nonlinear kernel NAP. Compared with linear NAP, nonlinear NAP with Gaussian kernel obtained about 11% relative improvement in equal error rate (EER).
Original languageEnglish
Title of host publication2008 IEEE International Conference on Acoustics, Speech and Signal Processing
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
Number of pages4
ISBN (Electronic)978-1-4244-1484-0
ISBN (Print)978-1-4244-1483-3
Publication statusPublished - 1 Mar 2008
Event2008 IEEE International Conference on Acoustics, Speech and Signal Processing - Caesars Palace, Las Vegas, NV, United States
Duration: 30 Mar 20084 Apr 2008


Conference2008 IEEE International Conference on Acoustics, Speech and Signal Processing
Country/TerritoryUnited States
CityLas Vegas, NV

Keywords / Materials (for Non-textual outputs)

  • Gaussian processes
  • polynomials
  • principal component analysis
  • speaker recognition
  • Gaussian kernels
  • SVM-based speaker verification systems
  • equal error rate
  • high-dimension feature space
  • kernel principal component analysis
  • nonlinear kernel
  • nonlinear kernel nuisance attribute projection
  • polynomial kernels
  • speaker verification
  • Input variables
  • Kernel
  • NIST
  • Polynomials
  • Principal component analysis
  • Research and development
  • Speaker recognition
  • Support vector machine classification
  • Support vector machines
  • Telecommunications
  • nuisance attribute projection
  • speaker location
  • supporting vector machines


Dive into the research topics of 'Nonlinear kernel nuisance attribute projection for speaker verification'. Together they form a unique fingerprint.

Cite this