Survey on language processing algorithm for sign language synthesizer

• Main Authors: Maarif, Haris Al Qodri, Akmeliawati, Rini, Gunawan, Teddy Surya
• Format: Article
• Language: English
• Published: Faculty of Art Design and Technology, University of Derby 2017
• Subjects: TK7800 Electronics. Computer engineering. Computer hardware. Photoelectronic devices
• Online Access: http://irep.iium.edu.my/72336/1/Maarif2017_Survey%20on%20Language%20Processing%20Algorithm%20for%20Sign%20Language%20Synthesizer.pdf; http://irep.iium.edu.my/72336/; http://ojs.unsysdigital.com/index.php/ijrm/article/view/10.21535%252Fijrm.v4i2.1001

Hearing and Speech Impaired (HSI) community is often disadvantaged due to their difficulties in communicating with the rest of the World. Although Sign languages have been used as a means to convey their message, problems still exist as there are not many people familiar with this type of languages. Besides, the number of available translators is insufficient to resolve the issue. This has motivated various research described in this paper. In general, we can divide the main contribution in solving the communication problem of the HSI community into two types, the first involves the development of automatic SL translators which allow the non-HSI communities to understand the messages conveyed by the HSI people, and the second relates to the development of SL synthesizers which allow the non-HSI community to pass their messages to HSI people. Thus, when both systems are integrated, two way communication can be established. In this paper, we provide a comprehensive review on the work in SL synthesizer development. SL synthesizer is a tool that synthesizes or constructs series of signs of a particular SL based on the speech input given to the system. The process involves adjusting speech input that can fit the sign language grammatical rule in order to provide an understandable output to the HSI people. Some algorithms, such as Natural Language Processing, are explored. In this paper, the details of each methods have been identified and discussed. In conclusion, we found that the Natural Language Processing (NLP) has provided relatively more efficient process and better results.