Automatic Gesture Analysis Using Constant Affine Velocity - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2014

Automatic Gesture Analysis Using Constant Affine Velocity

Résumé

Hand human gesture recognition has been an important research topic widely studied around the world, as this field offers the ability to identify, recognize, and analyze human gestures in order to control devices or to interact with computer interfaces. In particular, in medical training, this approach is an important tool that can be used to obtain an objective evaluation of a procedure performance. In this paper, some obstetrical gestures, acquired by a forceps, were studied with the hypothesis that, as the scribbling and drawing movements, they obey the one-sixth power law, an empirical relationship which connects path curvature, torsion, and euclidean velocity. Our results show that obstetrical gestures have a constant affine velocity, which is different for each type of gesture and based on this idea this quantity is proposed as an appropriate classification feature in the hand human gesture recognition field.
Fichier non déposé

Dates et versions

hal-01076675 , version 1 (22-10-2014)

Identifiants

Citer

Jenny Alexandra Cifuentes-Quintero, Pierre Boulanger, Minh Tu Pham, Richard Moreau, Flavio Prieto. Automatic Gesture Analysis Using Constant Affine Velocity. 36th IEEE EMBC, Aug 2014, Chicago, IL, United States. ⟨10.1109/EMBC.2014.6943964⟩. ⟨hal-01076675⟩
170 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More