The Virtual Digital Sculpture Based on Natural Gesture

the paper presents a new method of 3D modeling which copies our natural gesture to model. In the proposed method, a Leap Motion is fixed on the table to measure the position and orientation of the hands. The measurement errors will increase over time because of the tracking errors and the noise of equipment. Therefore, we use Kalman Filter and Particle Filter to estimate the position and orientation of our hands. 3D Max is used in our method to build the joint model of our hands. And then we use OSG (OpenSceneGraph) to refactor the model of our hands. When the gesture model and the sculpture model collide, we need to eliminate the common part of the collision, making the sculpture model to be deformed. And a triangle collision detection algorithm is used to achieve collision detection of 3D models. The proposed method was experimentally tested in a lab environment, and the results indicate that an operator can easily make a virtual sculpture using his/her 3D gesture even the operator is not an expert.