@inproceedings{camgoz-etal-2016-bosphorussign:lrec, author = {Camg{\"o}z, Necati Cihan and K{\i}nd{\i}ro{\u g}lu, Ahmet Alp and Karab{\"u}kl{\"u}, Serpil and Kelepir, Meltem and {\"O}zsoy, Ay{\c s}e Sumru and Akarun, Lale}, title = {{B}osphorus{S}ign: A {T}urkish {S}ign {L}anguage Recognition Corpus in Health and Finance Domains}, pages = {1383--1388}, editor = {Calzolari, Nicoletta and Choukri, Khalid and Declerck, Thierry and Goggi, Sara and Grobelnik, Marko and Maegaard, Bente and Mariani, Joseph and Mazo, H{\'e}l{\`e}ne and Moreno, Asuncion and Odijk, Jan and Piperidis, Stelios}, booktitle = {10th International Conference on Language Resources and Evaluation ({LREC} 2016)}, publisher = {{European Language Resources Association (ELRA)}}, address = {Portoro{\v z}, Slovenia}, day = {23--28}, month = may, year = {2016}, isbn = {978-2-9517408-9-1}, language = {english}, url = {https://aclanthology.org/L16-1220}, 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.} } @inproceedings{kubus:14026:sign-lang:lrec, author = {Kubu{\c s}, Okan}, title = {Discourse-based annotation of relative clause constructions in {Turkish} {Sign} {Language} ({TID}): A case study}, pages = {95--99}, editor = {Crasborn, Onno and Efthimiou, Eleni and Fotinea, Stavroula-Evita and Hanke, Thomas and Hochgesang, Julie A. and Kristoffersen, Jette and Mesch, Johanna}, booktitle = {Proceedings of the {LREC2014} 6th Workshop on the Representation and Processing of Sign Languages: Beyond the Manual Channel}, maintitle = {9th International Conference on Language Resources and Evaluation ({LREC} 2014)}, publisher = {{European Language Resources Association (ELRA)}}, address = {Reykjavik, Iceland}, day = {31}, month = may, year = {2014}, language = {english}, url = {https://www.sign-lang.uni-hamburg.de/lrec/pub/14026.html}, abstract = {The functions of relative clause constructions (RCC) should be ideally analyzed at the discourse level, since the occurrence of RCCs can be explained by looking at interlocutors' use of grammatical and intonational means (cf. Fox and Thompson, 1990). To date, RCCs in sign language have been analyzed at the syntactic level with a special focus on cross-linguistic comparisons (see e.g. Pfau and Steinbach, 2005; Branchini and Donati, 2009). However, to our knowledge, there is no systematic corpus-based analysis of RCCs in sign languages so far. Since the elements of RCCs are mostly non-manual markers, it is often unclear how to capture and tag these elements together with the functions of RCCs. This question is discussed in light of corpus-based data from Turkish Sign Language. Following Biber et al. (2007), the corpus-based analysis of RCCs in TID follows the ``top-down'' approach. In spite of modality-specific issues, the steps in the process of annotation and identification of RCCs in TID fairly resemble this approach. The advantage of using these multiple steps is that the procedure not only captures the discourse functions of the RCCs but also identifies different strategies for creating RCCs based on linguistic forms.} } @inproceedings{zwitserlood:08027:sign-lang:lrec, author = {Zwitserlood, Inge and {\"O}zy{\"u}rek, Asli and Perniss, Pamela}, title = {Annotation of Sign and Gesture Cross-linguistically}, pages = {185--190}, editor = {Crasborn, Onno and Efthimiou, Eleni and Hanke, Thomas and Thoutenhoofd, Ernst D. and Zwitserlood, Inge}, booktitle = {Proceedings of the {LREC2008} 3rd Workshop on the Representation and Processing of Sign Languages: Construction and Exploitation of Sign Language Corpora}, maintitle = {6th International Conference on Language Resources and Evaluation ({LREC} 2008)}, publisher = {{European Language Resources Association (ELRA)}}, address = {Marrakech, Morocco}, day = {1}, month = jun, year = {2008}, language = {english}, url = {https://www.sign-lang.uni-hamburg.de/lrec/pub/08027.html}, abstract = {In a 5-year project, we compare expressions in the spatial domain between two sign languages (German Sign Language and Turkish Sign Language), the co-speech gestures accompanying two spoken languages (German and Turkish), and the pantomime-like structures used by hearing people (German and Turkish) asked to convey information without speaking. The aim is to discover the similarities and differences in the use of space in expressing referent location and motion between the sign languages, and between the signing, co-speech gesture and no-speech pantomime modes. To this end, we are building a large video corpus of task-related discourse data (about 90 minutes per 15 participants per condition). The data will be described using the IMDI metadata standard and linguistically annotated using ELAN. Parts of the data will be made accessible for research and educational purposes on the Browsable Corpus based at the MPI for Psycholinguistics. \par In this presentation, we report the annotation conventions we have been developing based on collected data. There are two levels of annotation: (i) a descriptive level where we gloss signs and gestures according to the movements/positions of the hands, head, face, and body; and (ii) an analytic/coding level where each sign or gesture is analyzed with respect to the function of establishing and/or maintaining reference in discourse (e.g. through the use of pronouns, classifier predicates, modified verbs, and role shift in signing, and similar forms in gestures). Our conventions combine aspects from other annotation and coding systems developed for sign and gesture (e.g. the ECHO, Corpus NGT, and Auslan Corpus conventions; HamNoSys; gesture coding conventions as developed by Kita, Van Gijn and Van der Hulst), but go beyond them in placing special emphasis on coding both systems with the same parameters. \par On the descriptive level, we developed a 3-dimensional scheme to identify for hand orientation, location, and direction of signs and gestures, allowing comparison across languages. On the analytic/coding level, we devised ways of categorizing how the various spatial expressions in sign and gesture map onto different coreference devices in discourse. \par Parts of the sign and gesture data have now been annotated. We will present some generalizations and conclusions drawn from using our annotation conventions regarding cross-linguistic and sign-gesture comparison. Furthermore, based on our annotation experiences, we will discuss the advantages as well as the shortcomings of our annotation scheme and suggest specific improvements, which the linguistic community needs to consider in terms of ways they can be implemented in the technology of annotation software (such as ELAN).} }