A simplification algorithm based on appearance maintenance

This paper present a new simplification algorithm named EQEM which is based on the QEM algorithm to simplify geometry model with texture. The algorithm used a new framework to integrate geometry and texture factor into simplification process. In the framework, an error metric descriptor is described in detail. Firstly, Gauss curvature of the vertex should be calculated to assure good geometry similarity. Then, the descriptor take into account the visual importance in the simplification. In this way, we introduce a method of edge detection algorithm in image processing, which use mallat wavelet method to quickly extract the distinct edges from a texture image. Also, we compute the color distance of the vertices which are not belonged to any edges as the supplement of the metric descriptor. Experiment prove the simplification can make the sharp-simplified model keep similar appearance. For the descriptor is extended from the QEM error metric and can also be calculated by matrix multiplication, the algorithm is with high efficiency. After all, for all the parameters in the error metric formula are adjustable, algorithm can fit for different type of the mesh models and simplification scale.