Voice-driven applications: Architecture proposal and performance evaluation

Voice modality has a great potential to enrich standard WWW user experience, offer higher efficiency and overcome limits of visually impaired users. However, creating voice driven applications is currently complicated due to requirements of specific software and skills. To overcome this, we propose a novel architecture for large scale development and deployment of voice driven applications. Another contribution of this paper is performance evaluation to verify usability of voice modality in home environment. Various kinds of background noise were analyzed in order to determine possible drawbacks of speech recognition. Results were investigated and discussed to find out possible causes and further improvements of voice driven applications.