RT Journal Article
T1 Fast gesture recognition based on a two-level representation.
A1 Bandera-Rubio, Juan Pedro
A1 Marfil-Robles, Rebeca
A1 Bandera-Rubio, Antonio Jesús
A1 Rodríguez-Fernández, Juan Antonio
A1 Molina-Tanco, Luis
A1 Sandoval-Hernández, Francisco
K1 Interacción hombre-ordenador
K1 Análisis de series temporales
AB Towards developing an interface for human–robot interaction, this paper proposes atwo-level approach to recognise gestures which are composed of trajectories followedby different body parts. In a first level, individual trajectories are described by aset of key-points. These points are chosen as the corners of the curvature functionassociated to the trajectory, which will be estimated using and adaptive, non-iterativescheme. This adaptive representation allows removing noise while preserving detail incurvature at different scales. In a second level, gestures are characterised through globalproperties of the trajectories that compose them. Gesture recognition is performedusing a confidence value that integrates both levels. Experimental results show thatthe performance of the proposed method is high in terms of computational cost andmemory consumption, and gesture recognition ability.
PB Elsevier
YR 2009
FD 2009-10-01
LK https://hdl.handle.net/10630/34812
UL https://hdl.handle.net/10630/34812
LA eng
NO Bandera, J.P., Marfil, R., Bandera, A., Rodríguez, J.A., Molina-Tanco, L., Sandoval, F. FAST GESTURE RECOGNITION BASED ON A TWO-LEVEL REPRESENTATION Pattern Recognition Letters, 2009, 30(13), pp. 1181–1189
NO https://v2.sherpa.ac.uk/id/publication/11448
DS RIUMA. Repositorio Institucional de la Universidad de Málaga
RD 22 ene 2026