The rapid expansion of the mobile device market is driving new demand for cloud-based Automatic Speech Recognition (ASR) for speech transcription, applications supporting voice command and control, voice search, voice mail transcription and speech translation. The focus of the technology is hence to offer high accuracy general purpose speech recognition supporting Chinese, English and Asian languages.
The Cloud-based ASR allows end user to efficiently connect to the designated server while continuously performing the speech recognition. The decoded transcription is instantly returned after each utterance is being recorded and streamed to the server. The server can support multiple clients simultaneoulsy with different targetted languages.
The cloud-based ASR engine adopts fast and highly accurate speech decoding algorithm using the state-of-the-art weighted finite state transducers. Cloud-based (or server-client) architecture allows continual improvement on back-end algorithms and models, without affecting the installed user-base, leading to the increase in its accuracy. The multithreaded implementation provides the flexibility of scaling according to the size of the deployment. Other innovations include:
The cloud-based ASR APIs are easy to use and may be integrated with other applications, and hence, industrial partners can easily develop potential applications such as:
Contact us now to find out how our state-of-the-art ASR can help you to easily build and launch your innovative voice recognition applications.