Integrating HMM-Based Speech Recognition With Direct Manipulation In A Multimodal Korean Natural Language Interface

摘要

This paper presents a HMM-based speech recognition engine and its integrationinto direct manipulation interfaces for Korean document editor. Speechrecognition can reduce typical tedious and repetitive actions which areinevitable in standard GUIs (graphic user interfaces). Our system consists ofgeneral speech recognition engine called ABrain {Auditory Brain} and speechcommandable document editor called SHE {Simple Hearing Editor}. ABrain is aphoneme-based speech recognition engine which shows up to 97% of discretecommand recognition rate. SHE is a EuroBridge widget-based document editor thatsupports speech commands as well as direct manipulation interfaces.