An integrated approach to feature compensation combining particle filters and hidden Markov models for robust speech recognition

Mushtaq, Aleem; Hui-Lee, Chin

doi:10.1109/icassp.2012.6288982

Cited by 12 publications

(8 citation statements)

References 46 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Context‐based prediction can be used to determine the probability of a value, as the frequency of its occurrence in a certain context and, thus, it has been successfully applied as statistical model in several computer science areas such as computational biology [29], web mining [30], ubiquitous computing [31], information retrieval [32], speech recognition [33] and even in computer architecture [34]. Similar to a Markov process, it consists of a set of N distinct states S = { S 1 , S 2 , …, S N } [35].…”

Section: Description Of the Proposed Context‐based Prediction Filtementioning

confidence: 99%

Context‐based prediction filtering of impulse noise images

Gellért

Brad

2016

IET Image Processing

View full text Add to dashboard Cite

A new image denoising method for impulse noise in greyscale images using a context-based prediction scheme is presented. The algorithm replaces the noisy pixel with the value occurring with the highest frequency, in the same context as the replaceable pixel. Since it is a context-based technique, it preserves the details in the filtered images better than other methods. In the aim of validation, the authors have compared the proposed method with several existing denoising methods, many of them being outperformed by the proposed filter.

show abstract

Section: Description Of the Proposed Context‐based Prediction Filtementioning

confidence: 99%

Context‐based prediction filtering of impulse noise images

Gellért

Brad

2016

IET Image Processing

View full text Add to dashboard Cite

show abstract

“…Being a Monte Carlo method, particle filters are versatile and can handle a broad category of dynamical systems as they are not constrained by linearity and Gaussianity requirements that inhibit Kalman Filter [11] and the extended Kalman Filter [12]. Particle filter compensation (PFC) [13] [14] algorithms compensate noisy speech features by directly tracking the clean speech features in the spectral domain. The recognition is performed on melfrequency cepstral coefficient (MFCC) features extracted from the newly estimated filter bank features.…”

Section: Introductionmentioning

confidence: 99%

Joint tracking of clean speech and noise using HMMs and particle filters for robust speech recognition

Mushtaq

Lee

2012

2012 Conference Record of the Forty Sixth Asilomar Conference on Signals, Systems and Computers (ASILOMAR)

Self Cite

View full text Add to dashboard Cite

We propose a dynamic joint tracking framework to monitor the clean speech signal and noise simultaneously in order to compensate the noisy features. The clean speech signal is tracked using an integrated algorithm based on both particle filters and hidden Markov models. The information available from speech tracking is used for tracking and estimating the noise parameters. The availability of dynamic noise information enhances the robustness of the algorithm in case of large fluctuations in noise. We report on experimental results obtained with the Aurora-2 connected digit recognition task, and show that the performance for the additive noise cases can be improved by 12.15% over the state-of-the-art multi-condition training if the noise mean is updated every 300 milliseconds.

show abstract

“…We incorporate statistical information available in the acoustic models of clean speech, e.g., the HMMs trained with clean speech, as an alternative state transition model [10][11]. The similarity between HMMs and particles filters can be seen from the fact that an observation probability density function corresponding to each state of an HMM describes, in statistical terms, the characteristics of the source generating a signal of interest if the source is in that particular state, whereas in particle filters we try to estimate the probability distribution of the state the system is in when it generates the observed signal of interest.…”

Section: Speech and Noise Tracking For Noise Compensationmentioning

confidence: 99%