Modeling NERFs for speaker recognition

  title={Modeling NERFs for speaker recognition},
  author={Sachin S. Kajarekar and Luciana Ferrer and M. Kemal S{\"o}nmez and Jing Zheng and Elizabeth Shriberg and Andreas Stolcke},
We introduce a new type of feature to capture longrange patterns associated with individual speakers or wit h speaking styles. NERFs, or Nonuniform Extraction Region Feat ures, are defined based on regions of speech that are delimit d by various automatically extractable events of interes t. There is a wide unexplored space of potentially useful NERFs, but to use them successfully, at least two important challenge s must be addressed: (1) methods for coping with inherently m issing features… CONTINUE READING
Highly Cited
This paper has 25 citations. REVIEW CITATIONS


Publications citing this paper.
Showing 1-10 of 18 extracted citations

Similar Papers

Loading similar papers…