Reconstruction for shape feature representation in 3D human motion tracking
ZHOU Bin
MA Ling
Abstract：Aiming at the problem that during the reconstruction of shape feature in the 3D human body motion tracking, the reconstruction results of human motion postures compared with the real values have a large deviation and the effective tracking for a long term can not be realized, a reconstruction method for shape feature representation in 3D human body motion tracking based on the combined optimization method was proposed.The shape capture problem in 3D human body motion tracking was transformed into the regular programming problem with the POCS algorithm, and thus the shape capture problem in 3D human body motion tracking could be realized.The statistical estimation method of Bayes theory was adopted, and the captured shape mathematical modeling problem was transformed into the objective function maximization problem.With the Fisher linear discriminate analysis method, the shape feature points in 3D human body motion tracking were automatically located, and the located shape feature representation was reconstructed with the combined optimization method.The results show that the proposed method can effectively capture the shape feature in 3D human body motion tracking, and has high reconstruction precision and good accuracy.
Keywords：3D human bodymotion trackingshape featurerepresentationreconstructioncapturelocalizationPOCS algorithm
Publication Date：2017-01-01
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：6( 340-345 )