Premium
Human motion reconstruction from monocular images using genetic algorithms
Author(s) -
Zhao Jianhui,
Li Ling
Publication year - 2004
Publication title -
computer animation and virtual worlds
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.225
H-Index - 49
eISSN - 1546-427X
pISSN - 1546-4261
DOI - 10.1002/cav.44
Subject(s) - computer science , monocular , computer vision , artificial intelligence , motion (physics) , projection (relational algebra) , human motion , genetic algorithm , image (mathematics) , function (biology) , algorithm , machine learning , evolutionary biology , biology
This paper proposed an optimization approach for human motion recovery from the un‐calibrated monocular images containing unlimited human movements. A 3D skeleton human model based on anatomy knowledge is employed with encoded biomechanical constraints for the joints. Energy Function is defined to represent the deviations between projection features and extracted image features. Reconstruction procedure is developed to adjust joints and segments of the human body into their proper positions. Genetic Algorithms are adopted to find the optimal solution effectively in the high dimensional parameter space by simultaneously considering all the parameters of the human model. The experimental results are analysed by Deviation Penalty. Copyright © 2004 John Wiley & Sons, Ltd.