This work presents the models and methods used for the estimation of the joint center positions of posed skin meshes through the definition and repositioning of a personalized skeletal model. An articulated skeletal model is defined to represent the body skeleton using rigid segments connected by spherical joints. The model dimensions are personalized from a standing skin shape to estimate the joint center positions, then, the skeletal model is repositioned by inverse kinematics according to the skin shape in other postures to predict the new joint center positions. The results obtained from the skeletal model repositioning implicitly evaluated through the prediction of the posed skin shape using as a reference measured skin shapes for one person in various postures. The skin prediction errors obtained indicate that the models and methods are promising. For this reason, more efforts are needed to obtain ground truth data to validate the presented skeletal model and the method for its repositioning.