Computer Science – Sound
Scientific paper
2009-01-16
Computer Science
Sound
9 pages, 5 figures, Technical Report
Scientific paper
An effective way to increase the noise robustness of automatic speech recognition is to label noisy speech features as either reliable or unreliable (missing) prior to decoding, and to replace the missing ones by clean speech estimates. We present a novel method to obtain such clean speech estimates. Unlike previous imputation frameworks which work on a frame-by-frame basis, our method focuses on exploiting information from a large time-context. Using a sliding window approach, denoised speech representations are constructed using a sparse representation of the reliable features in an overcomplete basis of fixed-length exemplar fragments. We demonstrate the potential of our approach with experiments on the AURORA-2 connected digit database.
Cranen B.
Gemmeke J. F.
No associations
LandOfFree
TR01: Time-continuous Sparse Imputation does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with TR01: Time-continuous Sparse Imputation, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and TR01: Time-continuous Sparse Imputation will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-127131