Real-Time Bayesian Inference: A Soft Computing Approach to Environmental Learning for On-Line Robust Automatic Speech Recognition

Chowdhury, Foezur Rahman; Selouani, Sid‐Ahmed; O’Shaughnessy, D.

doi:10.1007/978-3-642-19644-7_47

Cited by 1 publication

(1 citation statement)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…9 Comparison of estimated noise spectrum of the proposed method (red line) with that of MCRA, MCRA2, and EMCRA the test case III at DFT frequency bin k = 75 (f = 750 Hz) and the results were published in Chowdhury et al (2011a). In this article we extend the performance analysis of the proposed on-line ASR in 8 more different types of nonstationary noises (e.g., car, exhibition hall, restaurant, street, airport, train station, subway MIR filtered and street MIR filtered) using Aurora 2 test data set 'a', set 'b', and set 'c'.…”

Section: Simulation Environments For On-line Asrmentioning

confidence: 99%

Bayesian on-line spectral change point detection: a soft computing approach for on-line ASR

Chowdhury

Selouani

O’Shaughnessy

2011

Int J Speech Technol

View full text Add to dashboard Cite

Current automatic speech recognition (ASR) works in off-line mode and needs prior knowledge of the stationary or quasi-stationary test conditions for expected word recognition accuracy. These requirements limit the application of ASR for real-world applications where test conditions are highly non-stationary and are not known a priori. This paper presents an innovative frame dynamic rapid adaptation and noise compensation technique for tracking highly non-stationary noises and its application for on-line ASR. The proposed algorithm is based on a soft computing model using Bayesian on-line inference for spectral change point detection (BOSCPD) in unknown nonstationary noises. BOSCPD is tested with the MCRA noise tracking technique for on-line rapid environmental change learning in different non-stationary noise scenarios. The test results show that the proposed BOSCPD technique reduces the delay in spectral change point detection significantly compared to the baseline MCRA and its derivatives. The proposed BOSCPD soft computing model is tested for joint additive and channel distortions compensation (JAC)-based on-line ASR in unknown test conditions using nonstationary noisy speech samples from the Aurora 2 speech database. The simulation results for the on-line AR show significant improvement in recognition accuracy compared to the baseline Aurora 2 distributed speech recognition (DSR) in batch-mode.Keywords On-line environment learning · Bayesian on-line inference for spectral change point detection · MCRA · On-line ASR · JAC compensation · Non-stationary noise tracking and estimate · Minimum search window · Frame dynamic · DSR · Highly non-stationary unknown test conditions · Real-world application · Smart phones and mobile hand-held devices · BOSCPD

show abstract

Section: Simulation Environments For On-line Asrmentioning

confidence: 99%