Factors that influence the performance of experienced speech recognition users
1Koester Performance Research, Ann Arbor, Michigan, USA.
Assistive Technology : the Official Journal of RESNA
|June 27, 2006
Summary
Effective correction strategies significantly improve automatic speech recognition (ASR) performance for individuals with physical disabilities. User computer time and typing speed also enhance ASR accuracy and text entry rates.
Area of Science:
- Assistive Technology
- Human-Computer Interaction
- Rehabilitation Engineering
Background:
- Automatic Speech Recognition (ASR) systems offer potential for individuals with physical disabilities.
- Performance variability among users with physical disabilities necessitates identifying key influencing factors.
Purpose of the Study:
- To identify factors affecting automatic speech recognition (ASR) performance in users with physical disabilities.
- To understand variations in recognition accuracy and text entry rates.
Main Methods:
- Bivariate and multivariate analyses were employed.
- Data collected from 23 experienced ASR users with physical disabilities.
- Evaluated the impact of 20 independent variables on ASR performance.
Main Results:
- Appropriate correction strategies demonstrated the strongest positive association with user performance.
- Increased computer usage time, higher manual typing speed, and faster ASR recognition speed correlated with better outcomes.
- ASR training amount and perceived adequacy did not significantly impact performance in this cohort.
Conclusions:
- Correction strategies are crucial for optimizing ASR effectiveness in users with physical disabilities.
- User-specific factors like typing speed and system efficiency play a significant role.
- Current ASR training approaches may not be the primary driver of performance gains for this group.
