3D human movement (walking) modeling using neural network

Computer graphics and animation has been widely used in many fields especially in film industry. The animation industry especially in Malaysia is limited to television commercial. This is because the production of animation film is very expensive. Furthermore the number of animators in Malaysia still needs to be increase. Modeling is the most important aspect in the animation development whereby articulated figure is commonly used in modeling the human or animal model. Kinematics technique is the common approach for control movement in animation. In this study, forward kinematics technique is applied into human animation and the defined angles are trained into neural network. The purpose of using neural network is to get the precise coordinate for “end effector”. The angles of the output are then applied to the model to produce human walking. This technique overcomes the problem of “try and error” in forward kinematics technique in order to get the target coordinate for human walking. This study is able to overcome the deficiency in forward kinematics. Therefore, forward kinematics can be use to produce an animation that involve a simple movement.