Hierarchical Mesh Segmentation Based on Quadric Surface Fitting

Abstract

In this paper we introduce a hierarchical region decomposition algorithm for triangle meshes segmentation based on quadric surface fitting. Initially, the whole mesh represents a single cluster. At every iteration, the region with largest error is divided and Lloyd iteration clustering is executed following, which make sure the segmentation has the minimum error in each hierarchical level. An error control strategy is used to get finally result automatically. Besides, users also can interpose this process by specifying the segmentation level. Each segmentation region after this process has a best fitting proxy of quadric surface. We propose an improved L 2,1 distance into a hybrid error, which is used in least square fit of quadric function. By comparing with the state-of-theart methods, our method is testified to be insensitive to noise. Model optimization and hole filling based on our algorithm are demonstrated in the last.

Feature-aligned segmentation using correlation clustering

“…Ideally, the feature lines emerge implicitly as the boundaries of the patches. The regularity condition can be low approximation error by an analytical primitive [5,[22][23][24][25], developability [26,27], or similarity using a variety of shape descriptors, including curvature [6,[28][29][30][31], normal voting tensor [32], slippage [33], and diffusion-type distances to a set of seed locations [34,35]. However, implicit methods can oversegment surface regions that are void of prominent feature lines, if the region fails the regularity conditions.…”