Are there any decent opensource projects out there (preferably with Python APIs) that do speaker or "voice recognition" reasonably well? I know this is an area of active research in academia.