Publication Details
Combination of strongly and weakly constrained recognizers for reliable detection of OOVs
Schwarz Petr, Ing., Ph.D. (DCGM FIT BUT)
Matějka Pavel, Ing., Ph.D. (DCGM FIT BUT)
Hannemann Mirko, Ph.D. (FIT BUT)
Rastrow Ariya (JHU)
White Christopher (JHU)
Khudanpur Sanjeev (JHU)
Heřmanský Hynek, prof. Ing., Dr.Eng. (DCGM FIT BUT)
Černocký Jan, prof. Dr. Ing. (DCGM FIT BUT)
speech processing, speech recognition, OOV detection, utterance verification
The paper is on the description of combination of strongly and weakly constrained recognizers for reliable detection of OOVs.
This paper addresses the detection of OOV segments in the output of large vocabulary continuous speech recognition (LVCSR) system. First, standard confidence measures based on frame-based word- and phone- posteriors are investigated. Substantial improvement was however obtained when posteriors from two systems - strongly constrained (LVCSR) and weakly constrained (phone posterior estimator) were combined. We show that this approach is suitable also for the detection of general recognition errors. All the results are presented on WSJ task with reduced recognition vocabulary.
@INPROCEEDINGS{FITPUB8494, author = "Luk\'{a}\v{s} Burget and Petr Schwarz and Pavel Mat\v{e}jka and Mirko Hannemann and Ariya Rastrow and Christopher White and Sanjeev Khudanpur and Hynek He\v{r}mansk\'{y} and Jan \v{C}ernock\'{y}", title = "Combination of strongly and weakly constrained recognizers for reliable detection of OOVs", pages = 4, booktitle = "Proc. International Conference on Acoustics, Speech, and Signal Processing (ICASSP)", year = 2008, location = "Las Vegas, US", publisher = "IEEE Signal Processing Society", ISBN = "1-4244-1484-9", language = "english", url = "https://www.fit.vut.cz/research/publication/8494" }