University of Twente Student Theses

Login

The adaptive presentation assistant using grammar-based recognition to support the process of presenting

Satink, Laurens (2009) The adaptive presentation assistant using grammar-based recognition to support the process of presenting.

[img] PDF
1MB
Abstract:Giving presentations has changed significantly over the last decades. Although the main reason why someone wants to give a presentation remains the same, the way it is done has changed significantly. During the last 30 years of the twentieth century, overhead slides displayed on an overhead projector were used for broadcasting the information. Over the past decades, these analogue overhead slides have been replaced by digital slides, projected by a beamer on a screen. More than the way of projecting, it is the slide itself that has changed teh most. Transparent overhead slides limited the information displayed to text, drawings, and with the possibility to photo copy on slides, figures and photo’s. Using digital slides made it possible to add enriched multimedia content, such as sound fragments, animations, (moving) pictures, movies, and appearing and vanishing text. These changes strongly influence the way a presentation can be given. In this thesis, we explore the possibility to apply speech and language technology (SLT) such as automated speech recognition (ASR) to allow the presenter to use (natural) speech to navigate through presentation. Several approaches were explored, developed and evaluated. First, a command-driven presentation assistant is created, after which titles and content of other slides are included. For each approach, a series of experiments is conducted and interpreted to evaluate the performance. The outcome shows that the presentation assistants in most cases are able to support the presenter in showing the correct slide at almost the correct time. A small but noticeable delay is inevitable when speech is used as a basis for the determination process, as the presenter discusses the next slide before the transition is initiated. The accuracy of the assistants is on average around 80-90% during conversational speech
Item Type:Essay (Master)
Faculty:EEMCS: Electrical Engineering, Mathematics and Computer Science
Subject:54 computer science
Programme:Interaction Technology MSc (60030)
Link to this item:https://purl.utwente.nl/essays/59217
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page