A two microphone noise reduction system is described. In an embodiment, input signals from each of the microphones are divided into subbands and each subband is then filtered independently to separate noise and desired signals and to suppress non-stationary and stationary noise. Filtering methods used include adaptive decorrelation filtering. A post-processing module using adaptive noise cancellation like filtering algorithms may be used to further suppress stationary and non-stationary noise in the output signals from the adaptive decorrelation filtering and a single microphone noise reduction algorithm may be used to further provide optimal stationary noise reduction performance of the system.

Two microphone noise reduction system

Modulation domain blind speech separation in noisy environments

We propose a novel approach to generate an ensemble of context-dependent deep neural networks (CD-DNNs) by using random forests of phonetic decision trees (RF-PDTs) and construct an ensemble acoustic model (EAM) accordingly for speech recognition. We present evaluation results on the TIMIT dataset and a telemedicine automatic captioning dataset and demonstrate the superior performance of the proposed RF-PDT+CD-DNN based EAM over the conventional CD-DNN based single acoustic model (SAM) in phone and word recognition accuracies.

Ensemble Acoustic Modeling for CD-DNN-HMM Using Random Forests of Phonetic Decision Trees

A analise das variaveis fisiologicas de um paciente pode melhorar a classificacao do risco
de obito de um paciente em uma Unidade de Terapia Intensiva(UTI) e auxiliar na tomada
de decisoes e alocacao de recursos disponiveis. Este trabalho propoe uma abordagem
computacional de analise de variaveis fisiologicas para previsao de obito de pacientes
em UTI. Variaveis fisiologicas que compoem series temporais(e.g., pressao arterial) sao
representadas como Processos Gaussianos Dependentes(PGDs). Variaveis que nao representam
series temporais(e.g., idade) sao utilizadas para agrupar os PGDs com Arvores
de Decisao. A classificacao e feita de acordo com uma medida de distância que combina
Deformacao Temporal Dinâmica e divergencia Kullback-Leibler. O resultado desta abordagem
quanto ao desempenho de classificacao e superior ao metodo padronizado SAPS-I
ja utilizado em UTI no conjunto de dados considerado para testes. O resultado e similar
a outros metodos computacionais publicados pela comunidade de pesquisa. Os resultados
comparando variacoes da abordagem proposta tambem mostram que ha vantagem em
utilizar o agrupamento de PGDs descrito.

/pdf/uma-abordagem-computacional-para-predicao-de-mortalidade-em-13mo1zysb6.pdf

Uma abordagem computacional para predição de mortalidade em utis baseada em agrupamento de processos gaussianos

Exemplar-based acoustic modeling is based on labeled training segments that are compared with the unseen test utterances with respect to a dissimilarity measure. Using a larger number of accurately labeled exemplars provides better generalization thus improved recognition performance which comes with increased computation and memory requirements. We have recently developed a noise robust exemplar matching-based automatic speech recognition system which uses a large number of undercomplete dictionaries containing speech exemplars of the same length and label to recognize noisy speech. In this work, we investigate several speech exemplar selection techniques proposed for undercomplete speech dictionaries to find a trade-off between the recognition accuracy and the acoustic model size in terms of the amount of speech exemplars used for recognition. The exemplar selection criterion has be to chosen carefully as the amount of redundancy in these dictionaries is very limited compared to overcomplete dictionaries containing plenty of exemplars. The recognition accuracies obtained on the small vocabulary track of the 2nd CHiME Challenge and the AURORA-2 database using the complete and pruned dictionaries are compared to investigate the performance of each selection criterion.

/pdf/data-selection-for-noise-robust-exemplar-matching-41da5tg0qj.pdf

Data selection for noise robust exemplar matching

We propose a novel approach of integrating exemplar-based template matching with statistical modeling to improve continuous speech recognition. We choose the template unit to be context-dependent phone segments (triphone context) and use multiple Gaussian mixture model (GMM) indices to represent each frame of speech templates. We investigate two different local distances, log likelihood ratio (LLR) and Kullback-Leibler (KL) divergence, for dynamic time warping (DTW)-based template matching. In order to reduce computation and storage complexities, we also propose two methods for template selection: minimum distance template selection (MDTS) and maximum likelihood template selection (MLTS). We further propose to fine tune the MLTS template representatives by using a GMM merging algorithm so that the GMMs can better represent the frames of the selected template representatives. Experimental results on the TIMIT phone recognition task and a large vocabulary continuous speech recognition (LVCSR) task of telehealth captioning demonstrated that the proposed approach of integrating template matching with statistical modeling significantly improved recognition accuracy over the hidden Markov modeling (HMM) baselines for both TIMIT and telehealth tasks. The template selection methods also provided significant accuracy gains over the HMM baseline while largely reducing the computation and storage complexities. When all templates or MDTS were used, using the LLR local distance gave better performance than the KL local distance. For MLTS and template compression, KL local distance gave better performance than the LLR local distance, and template compression further improved the recognition accuracy on top of MLTS while having less computational cost.

/pdf/integrated-exemplar-based-template-matching-and-statistical-3u67rac24y.pdf

Integrated exemplar-based template matching and statistical modeling for continuous speech recognition

We propose a novel approach to improve adaptive decorrelation filtering- (ADF-) based speech source separation in diffuse noise. The effects of noise on system adaptation and separation outputs are handled separately. First, fast noise compensation (NC) is developed for adaptation of separation filters, forcing ADF to focus on source separation; next, output noises are suppressed by speech enhancement. By tracking noise components in output cross-correlation functions, the bias effect of noise on the system adaptation objective function is compensated, and by adaptively estimating output noise autocorrelations, the speech separation output is enhanced. For fast noise compensation, a blockwise fast ADF (FADF) is implemented. Experiments were conducted on real and simulated diffuse noises. Speech mixtures were generated by convolving TIMIT speech sources with acoustic path impulse responses measured in a real room with reverberation time T60 = 0.3 second. The proposed techniques significantly improved separation performance and phone recognition accuracy of ADF outputs.

/pdf/fast-noise-compensation-and-adaptive-enhancement-for-speech-8slvb0kerr.pdf

Yunxin Zhao

Papers

Integrated exemplar-based template matching and statistical modeling for continuous speech recognition

Fast Noise Compensation and Adaptive Enhancement for Speech Separation