"Multilingual Automatic Detection of Structural Events in Speech"@en . " Mandarin" . "0"^^ . " automatically recognized text" . "V\u00EDcejazy\u010Dn\u00E1 automatick\u00E1 detekce struktur\u00E1ln\u00EDch ud\u00E1lost\u00ED v mluven\u00E9 \u0159e\u010Di" . . "automatic speech understanding" . . "2007-05-01+02:00"^^ . " Arabic" . . . "Tento projekt m\u00E1 za c\u00EDl podpo\u0159it u\u017E\u0161\u00ED spolupr\u00E1ci Katedry kybernetiky FAV Z\u010CU v Plzni se \u0161pi\u010Dkov\u00FDmi pracovi\u0161ti v USA, Human Language Technology Research Institute p\u0159i University of Texas at Dallas a v\u00FDzkumn\u00FDm institutem SRI International v oblasti automatick\u00E9ho rozpozn\u00E1v\u00E1n\u00ED \u0159e\u010Di. Projekt je zam\u011B\u0159en na automatickou detekci \u201Estruktur\u00E1ln\u00EDch ud\u00E1lost\u00ED\u201C v mluven\u00E9 \u0159e\u010Di, kter\u00E1 je kl\u00ED\u010Dov\u00E1 pro umo\u017En\u011Bn\u00ED n\u00E1sledn\u00E9ho automatick\u00E9ho zpracov\u00E1n\u00ED rozpoznan\u00E9ho textu. Mezi struktur\u00E1ln\u00ED ud\u00E1losti v \u0159e\u010Di pat\u0159\u00ED hranice v\u011Bt, \u201Eneplynulosti\u201C a dal\u0161\u00ED jevy, kter\u00E9 nejsou standardn\u011B ozna\u010Deny v \u201Eproudu slov\u201C, kter\u00FD je v\u00FDstupem konven\u010Dn\u00EDch automatick\u00FDch rozpozn\u00E1va\u010D\u016F \u0159e\u010Di. Hlavn\u00EDm c\u00EDlem spolupr\u00E1ce je vyv\u00EDjen\u00ED metod pro automatickou detekci struktur\u00E1ln\u00EDch ud\u00E1lost\u00ED ve zvukov\u00FDch dokumentech ve v\u00EDce r\u016Fzn\u00FDch jazyc\u00EDch, kter\u00E9 se od sebe v\u00FDrazn\u011B li\u0161\u00ED v syntaktick\u00E9, fonetick\u00E9 i prozodick\u00E9 struktu\u0159e. Pl\u00E1nov\u00E1na je pr\u00E1ce zejm\u00E9na na angli\u010Dtin\u011B a \u010De\u0161tin\u011B, po\u010D\u00EDt\u00E1 se ale i s roz\u0161\u00ED\u0159en\u00EDm pr\u00E1ce na arab\u0161tinu a mandar\u00EDnskou \u010D\u00EDn\u0161tinu, p\u0159\u00EDp. dal\u0161\u00ED jazyky." . "http://www.isvav.cz/projectDetail.do?rowId=ME 909"^^ . . . "automatic speech understanding; automatically recognized text; audio documents; Arabic; Mandarin; English"@en . "2010-06-30+02:00"^^ . . . "ME 909" . . . . . "12"^^ . "This project aims to support a closer cooperation between Department of Cybernetics, University of West Bohemia in Pilsen, the Human Language Technology Research Institute, University of Texas at Dallas, and SRI International, in the area of automatic speech understanding. In particular, the project is focused on automatic detection of structural events in speech, which is a key task for enabling downstream automatic processing of automatically recognized text. Structural events include sentence boundaries, disfluencies, and other phenomena that are currently not marked in the \u201Cstream of words\u201D output by conventional speech recognizers. The main goal is to develop methods for automatic detection of structural events in audio documents in different languages, including languages that differ significantly in phonetic, prosodic, and syntactic characteristics. We plan to mainly work on Czech and English, but we also plan to extend the work to Arabic and Mandarin, and eventually to other languages."@en . . . . "12"^^ . " audio documents" . "2009-02-24+01:00"^^ . "2009-12-31+01:00"^^ . "V projektu byla vytvo\u0159ena \u010Desk\u00E1 \u0159e\u010Dov\u00E1 datab\u00E1ze s anotac\u00ED tzv. struktur\u00E1ln\u00EDch metadat. Byl vyvinut automatick\u00FD syst\u00E9m pro detekci konce v\u011Bty v \u0159e\u010Di. \u00DAsp\u011B\u0161nost syst\u00E9mu byla analyzov\u00E1na z pohledu vlivu jazyka, \u017E\u00E1nru \u0159e\u010Di a identity \u0159e\u010Dn\u00EDka. Byla rozv\u00EDjena spolupr\u00E1ce Z\u010CU s UT Dallas a SRI Int."@cs . . . "0"^^ . "1"^^ . "We created a Czech speech database with structural metadata. We developed an automatic syst\u00E9m for sentence boundary detection. Syst\u00E9m performance has been analyzed in terms of language, genre, and speaker identity differences. The cooperation among UWB, UT Dallas and SRI Int. has been deepened."@en . .