N
Nam Soo Kim
Researcher at Seoul National University
Publications - 294
Citations - 4559
Nam Soo Kim is an academic researcher from Seoul National University. The author has contributed to research in topics: Speech enhancement & Hidden Markov model. The author has an hindex of 27, co-authored 275 publications receiving 4057 citations. Previous affiliations of Nam Soo Kim include KAIST & University College of Engineering.
Papers
More filters
Journal ArticleDOI
A statistical model-based voice activity detection
TL;DR: An effective hang-over scheme which considers the previous observations by a first-order Markov process modeling of speech occurrences is proposed which shows significantly better performances than the G.729B VAD in low signal-to-noise ratio (SNR) and vehicular noise environments.
Journal ArticleDOI
Voice activity detection based on multiple statistical models
TL;DR: This paper proposes a class of VAD algorithms based on several statistical models based on the Gaussian model, and incorporates the complex Laplacian and Gamma probability density functions to the analysis of statistical properties.
Journal ArticleDOI
Spectral enhancement based on global soft decision
Nam Soo Kim,Joon-Hyuk Chang +1 more
TL;DR: A novel speech enhancement technique based on global soft decision that provides a unified framework for such procedures as speech absence probability computation, spectral gain modification, and noise spectrum estimation using the same statistical model assumption.
Journal ArticleDOI
Speech recognition in noisy environments using first-order vector Taylor series
TL;DR: This paper generalizes relations between clean and noisy speech signal using vector Taylor series (VTS) expansion for noise-robust speech recognition and develops a detailed procedure to estimate environmental variables in the cepstral domain using the expectation and maximization algorithms based on the maximum likelihood (ML) sense.
Journal ArticleDOI
Statistical modeling of speech signals based on generalized gamma distribution
TL;DR: G/spl Gamma/D can model the distribution of the real speech signal more accurately than the conventional Gaussian, Laplacian, Gamma, or generalized Gaussian distribution (GGD).