Which approach to speech recognition avoids the problem caused by the variation in speech patterns among different speakers?

• Continuous speech recognition
• Isolated word recognition
• Connected word recognition
• Speaker-dependent recognition
• None of the above