Perera ILiyanapathirana LPMB202220222022Liyanapathirana, L.P.M.B. (2022). Architectural model for text to sign language interpreter (TSLI) [Master's theses, University of Moratuwa]. Institutional Repository University of Moratuwa. http://dl.lib.uom.lk/handle/123/22514http://dl.lib.uom.lk/handle/123/22514The deaf and hearing-impaired make up a considerable percentage of the world community having some essential needs that researchers have recently begun to target. There is no such proper way to communicate with deaf society instead of learning sign language. Sign language is a natural language that can be used as a powerful weapon for communicating with deaf society. The target objective of the research is to propose a Generic Architectural model to bridge the communication gap between deaf and non-deaf people. In the desertion, this problem was tackled by presenting the "Architectural model for Text to Sign Language Interpreter" system to translate text into any given Sign Language. Language parser Handler service is introduced as an API Contract. This Language parser Handler service acts as an abstraction layer for the backend services which are doing the NLP processing. Any researcher interested in translating sign language from a text can use this proposed model. A new language can be introduced to the model with minimum congurations. Instead of developing the entire model they just need to focus on the actual Language processing part. It should be a RESTFul microservice. This model can be used to test their API In a nutshell, the Language-specic parser service does the following. First, the individual words are extracted by tokenizing the input sentence. The morpho- logical analysis is then performed to identify the basic elements of the respective words. Basic components can be a sign stem, sign marker, or ngerspelling char- acters. After that, the identied basic word components (stream of tokens) are sent back to the client-side as a JSON string. The primary tokens are extracted from the JSON output. Then the tokens are mapped with the corresponding SiGML les which are forwarded to the signing avatar embedded in the client UI. The nal output is made by the signing avatar model which gives a stream of sign frames.enSIGN LANGUAGESERVICE ORIENTED ARCHITECTURECOMPUTER SCIENCE & ENGINEERING – DissertationCOMPUTER SCIENCE- DissertationThesis-AbstractEngineeringMSc in Computer Science & EngineeringDepartment of Computer Science & EngineeringTH4928