Résultats exploitables SPEECH RECOGNITION FOR DATA-ENTRY APPLICATIONS (SPEEDATA) Data-entry can be particularly costly when non electronic information - e.g. contained in documents or pictures - has to be interpreted by a domain expert before being stored on the computer - e.g. medical reporting, diagnostics, cataloguing, etc. SPEEDATA explores how to gain efficiency in this task by employing state-of-the-art ASR (Automatic Speech Recognition) technology. The ideal scenario is to let the computer work like a secretary who types what the expert user dictates, so that s/he does not have to worry about how data is entered and organised in the computer. Furthermore, the user can verify, correct and store the data through a user-friendly interface. SPEEDATA includes a continuous speech recogniser architecture specifically suited to data entry applications. It features real-time performance, keyword-optional data-entry recognition, and unsupervised speaker adaptation. The architecture is currently available for two operating system platforms: Unix and Windows. Two versions of the data-entry system are currently being tested in the field by two Italian public authorities. In both the tasks, a significant reduction in labour costs is expected. Efficiency will be gained in error reduction, less personnel required (as a typist is no longer needed) and data throughput. This is especially relevant where the amount of data to be input is vast. Project URL : http://poseidon.itc.it:7117/~voce/SPEEDATA/HomePage.html Recherche de données OpenAIRE... Une erreur s’est produite lors de la recherche de données OpenAIRE Aucun résultat disponible
