Theses in Process

Speech-to-speech translation systems - a critical analysis of existing solutions

Type:
Bachelor Thesis Business Information Systems
    Status:
    in process
    Tutor:

    Abstract

    Motivation:

    Speech translation is a process within which spoken phrases are instantly translated and spoken aloud in a second language. Speech translation systems, which typically integrate: speech recognition, machine translation and voice synthesis; enable speakers of different languages to communicate. It thus is of tremendous value for, among others, cross-cultural exchange and global business.

    Description:

    The main goal of this bachelor thesis is to critically analyse the maturity and applicability of existing speech-to-speech translation systems. Therefore, the scope of the thesis encompasses identification of applicable technologies, definition of a set of criteria that may be used to compare different solutions as well as identification of challenges that still need to be addressed. The conducted analysis should also allow to formulate a set of recommendations regarding the further development of the field.

    Exemplary literature:

    • Nakamura S (2009) Overcoming the language barriers with speech translation technology. Sci Technol Q Rev 31:35–48
    • Arora K, Arora S, Roy MK (2013) Speech to speech translation: a communication boon. CSI Trans ICT 1(3):207–213
    • Dutoit T (1997) An introduction to text-to-speech synthesis. Kulwer Academic Publishers
    • Guan Y, Zheng L, Tian J (2010) Real-time speaker adapted speech to speech translation system in mobile environment. In: 10th international conference on signal processing (ICSP). IEEE, pp 577–580
    • Nakamura S, Markov K, Nakaiwa H, Kikui G, Kawai H, Jitsuhiro T, Zhang JS (2006) The ATR multilingual speech to speech translation system. IEEE Trans Acoust Speech Signal Process 14(2):365–376
    • Zhou B, Cui X, Huang S, Cmejrek M, Zhang W (2013) The IBM speech-to-speech translation system for smartphone: Improvements for resource-constrained tasks. Comput Speech Lang 27(2):397–418