Optimizing recognition and rejection performance in wordspotting systems

Bourlard, Hervé; D'hoore, Bart; Boite, Jean-Marc

doi:10.1109/icassp.1994.389278

Cited by 71 publications

(34 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A garbage model of any random sequence of words or speech sounds is used to initialize lub . This can either be a separate garbage model evaluated for each frame or an estimate obtained from the local phone posterior probability estimates in a similar fashion to the "online garbage" approach in [31]. In the online garbage approach which was adopted for this work the most probable phone posteriors (excluding the most probable) are averaged and converted to a scaled likelihood by dividing by a uniform prior.…”

Section: B Pruningmentioning

confidence: 99%

Start-synchronous search for large vocabulary continuous speech recognition

Renals

Hochberg²

1999

IEEE Trans. Speech Audio Process.

View full text Add to dashboard Cite

Abstract-In this paper, we present a novel, efficient search strategy for large vocabulary continuous speech recognition. The search algorithm, based on a stack decoder framework, utilizes phone-level posterior probability estimates (produced by a connectionist/hidden Markov model acoustic model) as a basis for phone deactivation pruning-a highly efficient method of reducing the required computation. The single-pass algorithm is naturally factored into the time-asynchronous processing of the word sequence and the time-synchronous processing of the hidden Markov model state sequence. This enables the search to be decoupled from the language model while still maintaining the computational benefits of time-synchronous processing. The incorporation of the language model in the search is discussed and computationally cheap approximations to the full language model are introduced. Experiments were performed on the North American Business News task using a 60 000 word vocabulary and a trigram language model. Results indicate that the computational cost of the search may be reduced by more than a factor of 40 with a relative search error of less than 2% using the techniques discussed in the paper.Index Terms-Hidden Markov model, large vocabulary continuous speech recognition, phone deactivation pruning, search, stack decoding.

show abstract

Section: B Pruningmentioning

confidence: 99%

Start-synchronous search for large vocabulary continuous speech recognition

Renals

Hochberg²

1999

IEEE Trans. Speech Audio Process.

View full text Add to dashboard Cite

show abstract

“…This issue is typically addressed by using a more refined garbage model [6] or an on-line garbage model [7]. In this paper we propose to remove the phoneme models which are included in the keyword model from the filter model in the decoding network.…”

Section: Introductionmentioning

confidence: 99%

Keyword Detection for Spontaneous Speech

Liu

Billard

Bourlard

2009

2009 2nd International Congress on Image and Signal Processing

View full text Add to dashboard Cite

show abstract

“…The garbage modeling approach is employed for keyword spotting systems [1,5] and recently has also been applied to a continuous speech recognition task [6]. However, the detection performance for unknown words may be poor if the vocabulary size is large or when it is applied to a continuous speech task where the word boundaries are ambiguous.…”

Section: Introductionmentioning

confidence: 99%