Fundamentals of speech recognition

By L. Rabiner, Biing-Hwang Fred Juang. 1993

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Speech & Audio

Citations in the corpus

Transformer Transducer: A Streamable Speech Recognition Model with Transformer Encoders and RNN-T Loss (2020, methodology): “ is computationally intractable. Therefore, we deﬁne the forward variable (t;u) as the sum of probabilities for all paths ending at time-frame tand label position u. We then use the forward algorithm [14, 21] to compute the last alpha variable (T;U), which corresponds to P(yjx) deﬁned in Eq. (1). Ef”
Emotional speech recognition: Resources, features, and methods (2006, background): “2(a) (Rabiner and Juang, 1993).”
A Density Ratio Approach to Language Model Fusion in End-to-End Automatic Speech Recognition (2019, methodology): “-Backward algorithm can be used to calculate Eq. (7) efﬁciently during training, and Viterbi-based beam search (based on the argmax over possible alignments) can be used for decoding when Wis unknown [2, 28]. 3.3. Application of Shallow Fusion to RNN-T Shallow Fusion (Eq. (3)) can be implemented in ”
Temporal classification of multichannel near-infrared spectroscopy signals of motor imagery for developing a brain-computer interface (2007, methodology): “To recognize an unknown trial data, the likelihood of each model generating the trial data (observation vector) is calculated using the Viterbi algorithm (Rabiner, 1989, Rabiner and Juang, 1993), and the most likely model identifies the data as resulting from left-hand or right-hand task.”
A survey of temporal data mining (2006, methodology): “two sequences. Time warping methods have been used for sequence classification and matching for many years (Kruskal 1983; Juang & Rabiner 1993, chapter 4; Gold & Morgan 2000).”
Adversarial Attacks Against Automatic Speech Recognition Systems via Psychoacoustic Hiding (2018, background): “he number of necessary parameters is signiﬁcantly larger than that of the previous state-of-the-art Gaussian-MixtureModel probability densities within Hidden Markov Models (so-called GMM-HMM systems) [41]. As a consequence, this high number of parameters gives an adversary much space to explore (and”
Estimating deformations of isotropic Gaussian random fields on the plane (2008, methodology): “the deformation f is used in speech recognition to model local compression or expansion of time in different utterances of a spoken word (see [ 17 ]).”
Comparing passwords, tokens, and biometrics for user authentication (2003, methodology): “It is equivalent to applying two speech processing methods to authentication: verbal information verification [50] to verify that the speech-recognized result is the same as the challenge and speaker verification [51] to verify that the voice characteristics of the response are close to the user’s t”
Dimensionality Reduction of a Pathological Voice Quality Assessment System Based on Gaussian Mixture Models and Short-Term Cepstral Parameters (2006, methodology): “The suggested formula that models this relationship is as follows [ 27 ]:”
Audio-Visual Speech Modeling for Continuous Speech Recognition (2000, methodology): “2) Composite Model Formulation: Recognition using the Viterbi approximation [30] appears to be a continuous speech decoding problem where all of the concurrent word segmentations, as well as all of the HMM state segmentations, must be hypothesized.”
Facial expression recognition from video sequences: temporal and static modeling (2003, methodology): “common application of HMM is in speech recognition [ 33 ].”