Reconstruction Of Virtual Human And Animated Algorithm

Virtual human model is a kind of clothing-oriented model in garment manufacturing industry. On the one hand, it needs to meet the request of clothes cutting process. On the other hand, it needs to meet the request of users to reconstruct correct human model. This paper will base on these two aspects to describe the generation of virtual model from parameterized human modeling, modeling from photos and animation respectively.Parameterized human modeling is to construct 3D human model according to numerable anthropometry parameters input by users. Our work focus on two points: minimize user input and generate an accurate model for specific human body. The most difference with previous method is a model database is not necessary; every part of human body is independent. From the original model, we can get any customized human model by only nine parameters.We use four two-dimensional photos to realize the algorithm of reconstruction human model from photos. Our work mainly concentrates in searching feature points from 2D photos accurately and translating these feature points to deformation parameters. Our method has three characteristics: except requirements of size and background there are no other shooting requirements about photo; second, the feature pointes searching method base on the theory of Ergonomics to reduce time consuming and increase efficiency; third, this method is entirely automatic algorithm, there is no user interaction between photos imputing and deformation. Finally, the employment of efficient 3D rendering improves the quality of model.Animation will be more conducive to the model's show ability. This method based on the theory of key frame in traditional animation to adjust model movement, and then connect these rendering scenes together and play. The works mainly focus on the decomposition and scheduling in one animation. This paper will take walk as an example to describe the details about animation production.