【摘要】 人的视觉系统通常对模型的高曲率变化区域、尖角、边界轮廓等几何特征格外敏感。基于视觉特征度量的模型简化算法,在简化过程中通过调整网格顶点的简化误差来改变边折叠的次序,优先保留模型视觉敏感区域的网格;从而解决了模型网格的各部分简化力度趋于均衡的问题,最大限度地保留了原模型的视觉冲击。经实验,在同等简化力度下得到的简化结果能够突出原模型的视觉特征,具有明显优势。更多还原

【Abstract】 Some geometric features such as high-curvature regions, boundaries, corners and silhouette in low-resolution 3D models are sensitive to human. A new model simplification algorithm based on the visual feathers measurement was proposed which could preserve the sensitive geometry features first by changing the order of edge collapse. Hence, the problem was overcome that all the parts in the simplified models tended to have homogeneous mesh density. And such simplified models can preserve the visual impact of original models as most as possible. In the experimentation ,even better simplified models were obtained which have more triangle faces at high-curvature regions, boundaries, corners and silhouette when the simplification degree is the same to that of other algorithms.更多还原