Skip to Main content Skip to Navigation
Theses

Reconnaissance automatique de la parole continue : compensation des bruits par transformation de la parole

Abstract : Performances of an automatic speech recognition system degrade when test and training conditions do not match. Classical Stochastic Matching (SM) method proposes an off-line estimation of a compensation function that maximizes the likelihood of the compensated speech, given the optimal sequence of models proposed by the recognition process. We developed a new frame-synchronous technic based on SM : compensation is performed in parallel with the recognition. This is suitable to cope with slowly varying noise. We proposed two additional versions of our approach: -a tree structure of transformations is used to build a state-dependant non-linear compensation function. This is motivated by the fact that similar observations will be affected similarly by the environment. -a surveillance process monitoring the fluctuations in the environment is used to trigger the reinitialisation of the compensation process. This enables our algorithm to cope with environments experiencing sudden occurrences of noise.
Document type :
Theses
Complete list of metadata

Cited literature [57 references]  Display  Hide  Download

https://hal.univ-lorraine.fr/tel-01748111
Contributor : Thèses Ul <>
Submitted on : Thursday, March 29, 2018 - 11:25:19 AM
Last modification on : Tuesday, April 24, 2018 - 1:53:20 PM
Long-term archiving on: : Friday, September 14, 2018 - 8:04:15 AM

File

SCD_T_2004_0175_BARREAUD.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : tel-01748111, version 1

Collections

Citation

Vincent Barreaud. Reconnaissance automatique de la parole continue : compensation des bruits par transformation de la parole. Autre [cs.OH]. Université Henri Poincaré - Nancy 1, 2004. Français. ⟨NNT : 2004NAN10175⟩. ⟨tel-01748111⟩

Share

Metrics

Record views

65

Files downloads

76