VOCALinc
Abstract
As part of a growing trend in biometric identification, speaker recognition is becoming increasingly important to the nation s forensic laboratories, intelligence agencies, and military branches. Voice, along with other biometrics, such as fingerprints and iris patterns, is a unique characteristic that can be used to verify or identify individuals. Until recently, only spectrogram-based speaker recognition techniques were available. These techniques rely on electronically recorded speech signals, graphically represented as a function of time and frequency (i.e., a spectrogram), that analysts must manually evaluate an error-prone, subjective, inconsistent, and time-consuming process. VOCALinc, a stand-alone automated speaker recognition software developed by MIT Lincoln Laboratory, overcomes many of the issues inherent in manual techniques. Incorporating recent advances in speaker recognition technology, the software consists of a suite of speaker recognition algorithms that conduct the speaker comparisons and a graphical user interface that allows analysts to specify known data about uploaded voice samples and apply the algorithm(s) of their choice.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 2014
- Accession Number
- ADA625747
Entities
People
- Joseph P. Campbell
- Pedro A. Torres-carrasquillo
Organizations
- Massachusetts Institute of Technology