Signing Avatar Motion: Combining Naturality and Anonymity

Félix Bigand (1, 2) , Annelies Braffort (1, 2) , E. Prigent (1) , Bastien Berret (3, 4)

1 LISN - Laboratoire Interdisciplinaire des Sciences du Numérique
2 CNRS - Centre National de la Recherche Scientifique
3 CIAMS - Complexité, Innovation, Activités Motrices et Sportives
4 IUF - Institut Universitaire de France

This paper deals with the movement of virtual signers, and more particularly with the notion of biological motion and the issue of anonymization. Animating an avatar from motion models as biological as possible is important to ensure perceptual acceptability. Recording real movements thanks to motion capture systems and mapping them to an avatar can ensure this acceptability. However, such accurate systems may convey extra-linguistic information such as the signer's identity. For some applications, it will be necessary to determine the parameters of the movement that carry identity in order to control them.

Keywords: Sign Language; Virtual Signer; Signing Avatar; Animation; Biological Motion; Identity; Anonymization

Domains: Engineering Sciences [physics]

Félix Bigand, Annelies Braffort, E. Prigent, Bastien Berret. Signing Avatar Motion: Combining Naturality and Anonymity. International Workshop on Sign Language Translation and Avatar Technology, Sep 2019, Hambourg, Germany. ⟨hal-02400929⟩