Discrete-utterance recognition with a fast match based on total data reduction

Simple item page
dc.contributor.authorNouza, Jan
dc.date.accessioned2016-07-08
dc.date.available2016-07-08
dc.date.issued1996
dc.description.abstractIn the paper, a two-level classification scheme applicable to practical discrete-utterance recognition systems is presented. Both the fast and fine match employ CDHMM whole-word models. The fast match is based on total data reduction, which includes both the minimalization of the acoustic data flow (the numbers of speech frames and features) and the reduction of the basic HMM parameters (the numbers of states and mixtures). The optimal choice of the fast match parameters is a subject of the procedure that aims at minimizing the total classification time while preserving the maximum available recognition accuracy. On a medium-size vocabulary task (121 city names) the fast match reduced recognition time to approx. 20% (compared with the original one-level system) with a negligible loss of accuracy. The time savings were even more considerable in case of a system with multi-mixture HMMs.en
dc.formattext
dc.identifier.scopus2-s2.0-0030355991
dc.identifier.urihttps://dspace.tul.cz/handle/15240/16587
dc.language.isoen
dc.publisherIEEE, Piscataway, NJ, United States
dc.relation.ispartofInternational Conference on Spoken Language Processing, ICSLP, Proceedings
dc.sourced-scopus
dc.titleDiscrete-utterance recognition with a fast match based on total data reductionen
dc.typeconferenceObject
local.accessopen access
local.citation.epage2110
local.citation.spage2107
local.departmentDepartment of Textile Materials
local.event.edate1996-10-06
local.event.locationPhiladelphia, PA, USA
local.event.sdate1996-10-03
local.event.titleProceedings of the 1996 International Conference on Spoken Language Processing, ICSLP. Part 1 (of 4)
local.facultyFaculty of Textile Engineering
local.fulltextyes
local.identifier.coden264
local.notenefunguje RIV
local.relation.volume4
Files
Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
2-s2.0-0030355991-o.pdf
Size:
695.71 KB
Format:
Adobe Portable Document Format
Description:
Článek
Download
Collections