The Sign Language Dataset Compendium



The Dicta-Sign-LSF-v2 is an extended version of the French Sign Language sub-corpus of the Dicta-Sign Corpus providing primary data (videos), elicitation data, annotation data and a related annotation guide, as well as preprocessed signer data including facial pose, upper body pose and hand shape estimates. It contains nine dialogue sessions with 18 signers of French Sign Language covering the topic of travel. The data was annotated in more detail and a convolutional-recurrent learning network was trained on the data, drawing on a compact and generalisable modeling of the signers to provide a baseline for the recognition of lexical signs and non-lexical structures. Dicta-Sign-LSF-v2 was built at the Laboratoire d'informatique pour la mécanique et les sciences de l'ingénieur (LIMSI) in 2020.

Language French Sign Language
Size 11 hours recorded, 35000 tokens annotated
Participants 16 participants
Metadata Format not available
Translation French, 11 hours (100%) translated
Annotation Fully annotated (100%)
Lexical sign, illustrative sign, pointing, tag, number, typing and gesture annotated
See Braffort (2019) for more detail
Data Format Open Pose, CNN models
Licence CC BY-NC-SA 4.0
Access Access to videos, elicitation material and partial annotations via Ortolang requires registration
Webpage Dataset:
Institution Laboratoire d'informatique pour la mécanique et les sciences de l'ingénieur (LIMSI)

Cite as

Identifiant ISLRN: 442-418-132-318-7

Valentin Belissen, Annelies Braffort, Michèle Gouiffès. Dicta-Sign-LSF-v2: Remake of a Continuous French Sign Language Dialogue Corpus and a First Baseline for Automatic Sign Language Processing. LREC 2020, 12th Conference on Language Resources and Evaluation, 2020, Marseille, France. [hal-02541792] Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur (LIMSI) (2022). Dicta-Sign-LSF-v2 [Corpus]. ORTOLANG (Open Resources and TOols for LANGuage) -, v1,

Common tasks used in this corpus

Task Subject areas
# recordings – open access 8
# recordings – restricted access 8
Data available

Articles mentioned above

This entry was last modified on 29 June 2023.