Developing a video annotation tool for building Turkish sign language corpus: TIDVA

In Millions of people around the world suffer from hearing disability. This large number denotes the importance of developing sign language studies to allow hearing impaired people to communicate with others easily. One of the biggest problems in studying in the field of sign language is the absence of corpus. Developing corpus contributes the studies in the field of sign language by resolving this deficiency. In this thesis, a video annotation tool called TİDVA and Turkish Sign Language Corpus which was built by using TİDVA has been proposed. This thesis provides a great contribution to the field of Turkish Sign Language. Because TİDVA is a video annotation tool which is a unique study in the field of Turkish Sign Language. Moreover developed corpus eliminates lack of corpus in Turkish Sign Language. TİDVA is annotates videos, divides them into sub-videos and generates output to build corpus. 500 videos of sample sentences in the Turkish Sign Language Dictionary, which was developed by Ministry of Family, Labor and Social Services, were selected and then annotated by using TİDVA. The developed corpus has totally 596 sentences and 5343 words. Searching Tool was developed to make queries on the corpus and to list the results of the queries. Statistical analysis was made and results of this analysis were demonstrated. The currently developed corpus will contribute to future studies. It can be made richer by using the developed tool in the future. Keywords: Corpus, Turkish Sign Language, annotation tool, Turkish Sign Language corpus.