The Sign Language Dataset Compendium


Corpus

Dicta-Sign Corpus

The Dicta-Sign Corpus is a multilingual corpus of the four SLssign languages British Sign Language, German Sign Language, Greek Sign Language and French Sign Language covering the topic of European travel. The corpus collects at least 14 informants per language. The recording sessions took approximately two hours and the same elicitation materials were used across languages. Dicta-Sign was a three-years project from the European's seventh framework programme. The consortium conducting the project consisted of eight partners: Institute for Language and Speech Processing, Universität Hamburg, University of East Anglia, University of Surrey, Laboratoire d'informatique pour la mécanique et les sciences de l'ingénieur (Limsi), Université Paul Sabatier, National Technical University of Athens, WebSourd.

In total 1000 concepts with a SLsign language equivalent in each language were collected (see Dicta-Sign Lexicon) as well as training data for isolated signs and elicitation material for corpus collections.

Recordings were made with seven cameras recording the signers from different perspectives: front, side and top view as well as additional stereo cameras.

Languages British Sign Language, German Sign Language, Greek Sign Language, French Sign Language
Size 25 hours recorded, 1000 types per language elicited
Participants ~60 participants
14–16 participants per language
Metadata Format IMDI
Translation English, 0.6–5.5 hours translated per language
Annotation Different bundles of annotation for different data sets containing segmentation, glosses, gaze, clause boundaries, HamNoSys
not available
Data Format iLex
Licence Individual license agreement for researchers
Access Open access to subset of videos and elicitation material
Webpage Project page: https://www.sign-lang.uni-hamburg.de/dicta-sign/portal/
Institution Institute for Language and Speech Processing, Universität Hamburg, University of East Anglia, University of Surrey, Laboratoire d'informatique pour la mécanique et les sciences de l'ingénieur (LIMSI), Université Paul Sabatier, National Technical University of Athens, WebSourd

Cite as

not available

Common tasks used in this corpus

Task Route description
Corpus Language British Sign Language
# recordings – open access 0
# recordings – restricted access 8
Data available none
Task Route description
Corpus Language German Sign Language
# recordings – open access 0
# recordings – restricted access 7
Data available none
Task Route description
Corpus Language Greek Sign Language
# recordings – open access 0
# recordings – restricted access 8
Data available none
Task Route description
Corpus Language French Sign Language
# recordings – open access 0
# recordings – restricted access 8
Data available none
Task Subject areas
Corpus Language British Sign Language
# recordings – open access 8
# recordings – restricted access 8
Data available https://www.sign-lang.uni-hamburg.de/dicta-sign/portal/task.html
Task Subject areas
Corpus Language German Sign Language
# recordings – open access 7
# recordings – restricted access 7
Data available https://www.sign-lang.uni-hamburg.de/dicta-sign/portal/task.html
Task Subject areas
Corpus Language Greek Sign Language
# recordings – open access 8
# recordings – restricted access 8
Data available https://www.sign-lang.uni-hamburg.de/dicta-sign/portal/task.html
Task Subject areas
Corpus Language French Sign Language
# recordings – open access 8
# recordings – restricted access 8
Data available https://www.sign-lang.uni-hamburg.de/dicta-sign/portal/task.html

This entry was last modified on 6 January 2023.