• Publications
  • Influence
The SuperSID project: exploiting high-level information for high-accuracy speaker recognition
TLDR
This paper provides an overview of the structure, data, task, tools, and accomplishments of the SuperSID project. Expand
Using prosodic and conversational features for high-performance speaker recognition: report from JHU WS'02
TLDR
We report here on work from the JHU 2002 Summer Workshop exploring a range of prosodic features, using as testbed the 2001 NIST Extended Data task. Expand
Conditional pronunciation modeling in speaker detection
TLDR
We present a conditional pronunciation modeling method for speaker detection task that does not rely on acoustic vectors. Expand
Maximum Mutual Information and Word Classes
Herein, I present some notes concerning implementation of now classical method of data clustering, called Maximum Mutual Information Clustering. It was introduced in [Mercer et al., 1992] in contextExpand
On Linear DETs
TLDR
We show that the linear DETs on a normal deviate scale are not exclusively produced by normal distributions, however, that normal distributions do play a unique role in the threshold behavior as one moves along the DET line. Expand
Optimal detection in case of the sparse training data
TLDR
We introduce an algorithm which minimizes the probability of classification error using only the independence assumption in the speaker detection task using discrete alphabet. Expand
New Methods in Statistical Speech Recognition