A feature-preserving simplification based on integral invariant clustering

Detailed models are required in computer graphics for many applications. However considering the processing and transporting time, it is often necessary to approximate these models. In this paper we provide an effective simplification method for mesh models, which decreases the size of complex models and keeps visual features. We employ the integral invariant to distinguish the desired features on the models with different scales, then use the k-means clustering algorithm to find the fixed feature vertex cluster in which the vertices are kept approximately identical by our best, finally provide a weighting map to guide the simplifications. The proposed algorithm by this paper provides significant improvement on feature-preserving, especially sharp feature-preserving, and it can also be combined with other mesh simplification schemes to improve their effects.