Human Body Modeling Based on Pose and Shape Space

In this paper, we propose a method to create a 3D model of a human body using commercial depth cameras. Due to the cheaply available commercial depth cameras, various systems that use 3D scan can now be realized at natural environment. However, due to the poor quality of raw depth data, instead of using the depth input directly to create a 3D model, we try to extract the pose and the shape of a person,and then use these information to deform a template model. We use a parametric body model called SCAPE. SCAPE body model employs a low dimensional model of shape and pose dependent deformation that is learnt from the database of range scans of human bodies. In our research, we try to estimate the parameters to model the human body using the Kinect depth information.In this paper, we propose a method to create a 3D model of a human body using commercial depth cameras. Due to the cheaply available commercial depth cameras, various systems that use 3D scan can now be realized at natural environment. However, due to the poor quality of raw depth data, instead of using the depth input directly to create a 3D model, we try to extract the pose and the shape of a person,and then use these information to deform a template model. We use a parametric body model called SCAPE. SCAPE body model employs a low dimensional model of shape and pose dependent deformation that is learnt from the database of range scans of human bodies. In our research, we try to estimate the parameters to model the human body using the Kinect depth information.