Towards robust and flexible continuous Putonghua recognition
Dr Huo, Qiang (Principal investigator)
RGC General Research Fund (GRF)
HKU Project Code
General Research Fund (GRF)
To design a task-independent Putonghua training database suitable for a wide range of vocabularies and applications; to choose a set of basic speech units based on the above task-independent database and how to train the corresponding HMMs to achieve a reasonable task-independent performance; to achieve high performance speech recognition through adaptive system design to quickly meet changing tasks, speakers and speaking environments.