An Out-of-Core Scheme for Simplification of Point-Sampled Models

Most of the exiting point-based simplification methods can´t cope with models that are too large to fit in main memory. We present an out-of-core scheme for simplifying large point-sampled models. Firstly, we obtain an initial simplified model using uniform clustering. Then we decimate the redundant details in the flat regions in the condition of keeping the maximum error less than the upper error limit. Finally the same number of details are used to refine the uneven regions. In order to preserve the boundaries of point-sampled models, we design an efficient scheme without using connectivity information. Experiment results show that our simplification method can simplify large models with high quality and low memory usage.