Adaptive multi‐scale analysis for point‐based surface editing

Abstract

Figure 1: Feature-based editing of a detailed point cloud (1.5 millions points). After a prior analysis of the input model (a) to detect, count and extract pertinent scales, the user can edit the geometry in real-time using a graphic equalizer to, for instance, remove the two first level of details (b), remove only the scratches and skin pores (c), or boost them and remove the wrinkles defined at an intermediate scale (d). AbstractThis paper presents a tool that enables the direct editing of surface features in large point-clouds or meshes. This is made possible by a novel multi-scale analysis of unstructured point-clouds that automatically extracts the number of relevant features together with their respective scale all over the surface. Then, combining this ingredient with an adequate multi-scale decomposition allows us to directly enhance or reduce each feature in an independent manner. Our feature extraction is based on the analysis of the scale-variations of locally fitted surface primitives combined with unsupervised learning techniques. Our tool may be applied either globally or locally, and millions of points are handled in real-time. The resulting system enables users to accurately edit complex geometries with minimal interaction.

Stable and efficient differential estimators on oriented point clouds

“…The APSS is a classical method developed in the context of point cloud rendering, but it is also used to analyse the point‐sampled surface as done by Mellado et al [MGB∗12] with the Growing Least Squares (GLS). They provide point‐wise multi‐scale differential descriptors based on the APSS that are widely used for point cloud registration [MDS15], modeling [NGM14], and pattern recognition [LMBM20, HLP∗20]. In addition to its efficiency and robustness, we rely on the APSS for the analytical formula of the sphere fitting, which makes it suitable for an integral invariant analysis and for kernel differentiation.…”