Formulation of the REMOS concept from an uncertainty decoding perspective

摘要: In this paper, we introduce a new formulation of the REMOS (REverberation MOdeling for Speech recognition) concept from an uncertainty decoding perspective. Based on convolutive observation model that relaxes conditional independence assumption hidden Markov models, effectively adapts automatic speech recognition (ASR) systems to noisy and strongly reverberant environments. While approaches are typically designed operate irrespectively employed routine ASR system, explicitly considers additional information provided by Viterbi decoder. contrast previous publications concept, provide conclusive derivation its using Bayesian network representation in order prove inherent character.