. "LIVE TV SUBTITLING - Fast 2-pass LVCSR System for Online Subtitling" . "RIV/49777513:23520/07:00000324" . . . "RIV/49777513:23520/07:00000324!RIV08-MSM-23520___" . . "4"^^ . "\u010Cl\u00E1nek popisuje rychl\u00FD dvoupr\u016Fchodov\u00FD rozpozn\u00E1vac\u00ED syst\u00E9m spojit\u00E9 \u0159e\u010Di s velk\u00FDm slovn\u00EDkem (LVCSR) pro automatick\u00E9 online titulkov\u00E1n\u00ED \u017Eiv\u00FDch televizn\u00EDch po\u0159ad\u016F. Implementace navr\u017Een\u00E9ho syst\u00E9mu m\u016F\u017Ee b\u00FDt pou\u017Eita jak pro p\u0159\u00EDm\u00E9 rozpozn\u00E1v\u00E1n\u00ED audio kan\u00E1lu televizn\u00EDho programu, tak pro rozpozn\u00E1v\u00E1n\u00ED st\u00EDnov\u00E9ho \u0159e\u010Dn\u00EDka, kter\u00FD p\u0159emlouv\u00E1 origin\u00E1ln\u00ED zvukov\u00FD kan\u00E1l. Prvn\u00ED \u010D\u00E1st \u010Dl\u00E1nku se zam\u011B\u0159uje na p\u0159\u00EDpravu adaptivn\u00EDho jazykov\u00E9ho modelu pro televizn\u00ED programy, ve kter\u00FDch jsou osobn\u00ED jm\u00E9na specifick\u00E1 pro ka\u017Edou titulkovac\u00ED \u00FAlohu a musej\u00ED b\u00FDt p\u0159id\u00E1na do rozpozn\u00E1vac\u00EDho slovn\u00EDku. Druh\u00E1 \u010D\u00E1st nasti\u0148uje koncepci rozpozn\u00E1v"@cs . "The paper describes a fast 2-pass large vocabulary continuous speech recognition (LVCSR) system for automatic online subtitling of live TV programs. The proposed system implementation can be used for direct recognition of TV program audio channel or recognition of a shadow speaker who re-speaks the original audio channel. The first part of this paper focuses on preparation of an adaptive language model for TV programs, where person names are specific for each subtitling session and have to be added to the recognition vocabulary. The second part outlines the recognition system conception for automatic online subtitling with vocabulary up to 150 000 words in real-time. The recognition system is based on Hidden Markov Models, lexical trees and bigram and quadgram language models in the first and second pass, respectively. Finally, experimental results from our project with the Czech Television are reported and discussed."@en . "Lisabon" . "ASR; LVCSR; HMM; real-time; class-based language model; live TV; online subtitling"@en . . "INSTICC PRESS" . . "LIVE TV SUBTITLING - Fast 2-pass LVCSR System for Online Subtitling"@en . . "4"^^ . . "139-142" . "4"^^ . . . . . . "P(2C06020)" . "Barcelona" . "SIGMAP 2007" . "[AAACFFF74519]" . . . . "2007-01-01+01:00"^^ . . "978-989-8111-13-5" . "Titulkov\u00E1n\u00ED \u017Eiv\u00FDch televizn\u00EDch po\u0159ad\u016F - rychl\u00FD dvoupr\u016Fchodov\u00FD LVCSR syst\u00E9m pro online titulkov\u00E1n\u00ED"@cs . "431241" . "The paper describes a fast 2-pass large vocabulary continuous speech recognition (LVCSR) system for automatic online subtitling of live TV programs. The proposed system implementation can be used for direct recognition of TV program audio channel or recognition of a shadow speaker who re-speaks the original audio channel. The first part of this paper focuses on preparation of an adaptive language model for TV programs, where person names are specific for each subtitling session and have to be added to the recognition vocabulary. The second part outlines the recognition system conception for automatic online subtitling with vocabulary up to 150 000 words in real-time. The recognition system is based on Hidden Markov Models, lexical trees and bigram and quadgram language models in the first and second pass, respectively. Finally, experimental results from our project with the Czech Television are reported and discussed." . . . "LIVE TV SUBTITLING - Fast 2-pass LVCSR System for Online Subtitling" . "Titulkov\u00E1n\u00ED \u017Eiv\u00FDch televizn\u00EDch po\u0159ad\u016F - rychl\u00FD dvoupr\u016Fchodov\u00FD LVCSR syst\u00E9m pro online titulkov\u00E1n\u00ED"@cs . "Psutka, Josef" . "M\u00FCller, Lud\u011Bk" . "23520" . . "Pra\u017E\u00E1k, Ale\u0161" . . . "LIVE TV SUBTITLING - Fast 2-pass LVCSR System for Online Subtitling"@en . .