BosphorusSign: A Turkish Sign Language Recognition Corpus in Health and Finance Domains

BosphorusSign: A Turkish Sign Language Recognition Corpus in Health and Finance Domains Necati Cihan Camg¨oza, Ahmet Alp Kındıro˘glua, Serpil Karab¨ukl¨ub, Meltem Kelepirb, A. Sumru ¨Ozsoyb, Lale Akaruna aBo˘gazic¸i University, Computer Engineering Department, Istanbul, Turkey bBo˘gazic¸i University, Department of Linguistics, Istanbul, Turkey {cihan.camgoz, alp.kindiroglu, serpil.karabuklu, meltem.kelepir, ozsoys, akarun}@boun.edu.tr Abstract There are as many sign languages as there are deaf communities in the world. Linguists have been collecting corpora of different sign languages and annotating them extensively in order to study and understand their properties. On the other hand, the field of computer vision has approached the sign language recognition problem as a grand challenge and research efforts have intensified in the last 20 years. However, corpora collected for studying linguistic properties are often not suitable for sign language recognition as the statistical methods used in the field require large amounts of data. Recently, with the availability of inexpensive depth cameras, groups from the computer vision community have started collecting corpora with large number of repetitions for sign language recognition research. In this paper, we present the BosphorusSign Turkish Sign Language corpus, which consists of 855 sign and phrase samples from the health, finance and everyday life domains. The corpus is collected using the state-of-the-art Microsoft Kinect v2 depth sensor, and will be the first in this sign language research field. Furthermore, there will be annotations rendered by linguists so that the corpus will appeal both to the linguistic and sign language recognition research communities. Keywords: Turkish Sign Language (TID), Microsoft Kinect v2, Health, Finance, Sign Language Recognition 1. Introduction Sign Languages (SLs) are the natural means of commu- nication of deaf communities. Each deaf community has its own sign language with specific rules of its own. Al- though the deaf can communicate with each other via sign language, the communication between the hearing and the Deaf is an issue and often requires interpreters. This issue becomes more crucial when the need to communicate be- comes essential for a deaf person to fulfill their daily needs, such as getting services in a hospital or a bank. Turkish Sign Language (T¨urk ˙Is¸aret Dili, TID) is the lan- guage used by the deaf community of Turkey. The Turkish Deaf community also has trouble communicating with the hearing, as hearing community members rarely know Turk- ish Sign Language. Furthermore, literacy rates are low in the Deaf community, which also limits their communica- tion. Recently instated Turkish laws mandate the use of sign language interpreters for public services. However, there is not enough trained personnel to deploy in each of- fice and this absence has necessitated more feasible solu- tions such as automated translation services. Sign language linguists have been studying Turkish Sign Language in recent years by collecting large corpora and analyzing the aspects of the language ( ¨Ozsoy et al., 2013). However, there is no domain-specific corpus for Health and Finance. In this study, we are presenting Turkish Sign Lan- guage corpora in Health and Finance domains using a state- of-the-art depth sensor, Microsoft …

Content truncated to 500 words