Voice Server with Under-resourced Acoustic Model: Application to Agricultural Extension

Minh H. Le¹, Truong M. Tran², Nhut M. Pham³, Quan H. Vu⁴

^1,2,3,4Artificial Intelligence Laboratory, University of Science, VNU-HCM, HCM, Vietnam

Abstract

The majority of farmers in developing countries are unable to reach farming information and knowledge, and so they often rely on rudimentary methods. To enhance the agricultural life and productivity, one must ensure the flow of information to farmers; that is to bridge an agricultural info channel to farmers effectively. In this work, we move on to build an agricultural IVR system with a mission of providing agricultural info services to the farmers. However, since the underlying techniques resort to statistical approaches, a large amount of training material is demanded to reach a sustainable level of performance, especially on a new applied domain. Unfortunately, the required effort to develop such corpora is both costly and time consuming, and large scale acquisition campaigns might not be feasible. Under these restricted circumstances, subspace Gaussian mixture models and parameter synthesis are the keys to build ASR and TTS components of an IVR system. Experimental results confirm the hypothesis with 3.43% winning over conventional methods in an application of agricultural extension.

Keywords

Author's Profile

Author profile can be generated and linked through our partners World Book of Researchers. To include your profile online Click Here. After it is approved, please email to edlib @ asdf.res.in to create a link with all the papers.

• Minh H. Le : Profile
• Truong M. Tran : Profile
• Nhut M. Pham : Profile
• Quan H. Vu : Profile

e-AID

Cite this Article as Follows

Minh H. Le, Truong M. Tran, Nhut M. Pham, Quan H. Vu. Voice Server with Under-resourced Acoustic Model: Application to Agricultural Extension. International Conference on Cloud Computing and eGovernance 2014. Vol. 1. Chennai: Association of Scientists, Developers and Faculties, 2014. 114-120. Print.