Automatic Human Model Generation

摘要

The contribution presents an integrated system for automatic acquisition of a human torso model, using different input images. The output model consists of two free-form surface patches (with texture maps) for the torso and the arms. Also, the positions for the neck joint on the torso, and six joint positions on the arms (for the wrist, elbow and shoulder) are determined automatically. We present reconstruction results, and, as application, a simple tracking system for arm movements.