Mesh Simplification Method Research Based On Feature Preserved

With the rapid development of computer graphics, 3D models are becoming more common in people’s lives, and becoming another digital media gradually after voice, image, and video. It has been used widely in many field, like movie, games, animation, industrial design, and visualization. Especially in recent years, the rapid development of the internet brings more opportunities to graphics, playing an irreplaceable role in high-tech areas such as virtual reality and telemedicine. But with the continuous progress of the times, the amount of data is becoming more and more big, in this background, the progressing of the 3D model is facing an enormous challenge. How to deal with big mesh models is one of the difficult problem for graphics researchers. When we get a digital model for computer to processing of the object by scanning, we usually want to display every detail of the object as exhaustive as possible. But this would produce a vary lager mesh which contains hundreds thousands even millions faces, it will bring many problems when we process this so big mesh. So mesh model simplification is a necessary operation for these lager models.In the research filed of mesh simplification, there are many classic method, such as Lounsbery’s wavelet decomposition, Rossignax’s vertex cluster, Schroeder’s vertex delete, Hoppe’s edge collapse and Garland’s quadric error metric, they are all acquire approving result, but they are not take model details into account, so their result are all homogeneous and lost detail information of the model. In recent years, many mesh simplification methods based on preserving model significant features are proposed. Lee’s mesh saliency, Liu’s critical feature et al. These methods have same ideal, finding one measure standard to weigh mesh feature, so the significant features have lager value, these features will be preserved when simplified.We present a new simplification method based on this ideal too. First, calculate mesh saliency with Morse function and bilateral filter, and then combine to simplification method(QEM) to change the order of edge collapse. We consider mesh topology and local feature to weigh mesh saliency, the result show this approach is very effective.