I-Vector
Definition
A fixed-length low-dimensional speaker representation extracted by projecting a speaker's UBM statistics into a total variability space. Introduced by Dehak et al. (2011); the dominant speaker recognition paradigm from approximately 2011 to 2018.
- Introduced by
- Dehak et al., 2011
- Derived from
- UBM statistics projected into a total variability space
- Dominant era
- Approximately 2011 to 2018
- Superseded by
- x-vector, a deep neural network embedding
Common questions
Why was the i-vector approach considered a major advance in speaker recognition?+
It compressed both speaker and channel variability into a single fixed-length vector that could be compared with simple statistical scoring, replacing earlier, more computationally heavy joint factor analysis models with something practical for casework.
Why has i-vector been largely replaced by x-vector systems?+
Deep neural network embeddings trained on large speaker datasets generally achieve better discrimination, especially on short or noisy recordings, than the GMM-based statistics i-vectors rely on, though i-vector systems remain relevant in older validation studies and some legacy casework.
Related terms
- Background Population Database
- A corpus of speaker recordings used to calibrate the LR in a forensic comparison. Must be matched to the case in language,...
- BATVOX
- A commercial automated speaker recognition system (originally Agnitio, later Nuance Communications) widely used in operational forensic casework in Europe and Latin America....
- C_llr (Log-Likelihood Ratio Cost)
- A proper scoring rule measuring the calibration quality of a speaker recognition system's LR outputs; a well-calibrated system has C_llr close to...
- ENFSI BPM
- The ENFSI Best Practice Manual for Forensic Comparison of Speech (2015, rev. 2022); the operational standard for speaker comparison in European forensic...
- ENFSI Verbal Scale
- A standardised table published by the European Network of Forensic Science Institutes that maps ranges of LR values to verbal expressions of...
- Equal Error Rate (EER)
- The point on a classifier's detection error tradeoff curve where the false accept rate equals the false reject rate. Lower EER indicates...
- NIST SRE (Speaker Recognition Evaluation)
- A competitive benchmarking series run by the National Institute of Standards and Technology since 1996, in which research and commercial teams submit...
- PLDA (Probabilistic Linear Discriminant Analysis)
- A statistical back-end model used with i-vector and x-vector systems to compute a similarity score between two utterance representations, normalised for within-speaker...
- X-Vector
- A deep-learning speaker embedding produced by the penultimate layer of a time-delay neural network (TDNN) trained for speaker classification. Introduced by Snyder...