A comparison of hole-filling methods in 3D

Abstract

This paper presents a review of the most relevant current techniques that deal with hole-filling in 3D models. Contrary to earlier reports, which approach mesh repairing in a sparse and global manner, the objective of this review is twofold. First, a specific and comprehensive review of hole-filling techniques (as a relevant part in the field of mesh repairing) is carried out. We present a brief summary of each technique with attention paid to its algorithmic essence, main contributions and limitations. Second, a solid comparison between 34 methods is established. To do this, we define 19 possible meaningful features and properties that can be found in a generic hole-filling process. Then, we use these features to assess the virtues and deficiencies of the method and to build comparative tables. The purpose of this review is to make a comparative hole-filling state-of-the-art available to researchers, showing pros and cons in a common framework.

Reconstruction of the Sculpture Surface in the Process of Optical 3D Scanning

“…At the moment there is no general approach to solve this problem. Existing methods can be divided into three categories [5]: methods based on polygonal representation (restores arbitrary holes and irregular data can be used, but requires significant computational resources and user involvement in processing), methods based on parametric representation (does not require user involvement in the formation of the model but allows you to repair only small holes, requires holes to be regular and it does not work well with highly curved surfaces), methods based on signed distance functions and voluminous performances (restores arbitrary holes, irregular data can be used, restores large holes and does not require user input but requires significant computational resources and high resolution input). Professional software packages (RapidForm XOR, RapidForm XOR, etc.)…”

A Fast Hole-Filling Method for Triangular Mesh in Additive Repair

“…According to the work from Ju et al [16], the hole-filling methods based on a triangular mesh can be classified into two groups: surface-based methods and volume-based methods. In most cases, the former operates directly on the given mesh and follows the same stages: hole detection, initial filling and refinement [4]. The latter converts the input mesh into a signed distance function over a volumetric grid to fill holes and then extracts a completed mesh from a zero-level set of the distance function [17].…”

“…The main advantage of volume-based methods is its robustness in resolving geometric errors, and the drawback is the loss of geometric detail. For more works, please refer to these two survey papers [3,4]. Based on the above analysis, fast hole-filling, especially for different sizes of holes, is still a challenge.…”

“…Although many hole-filling techniques have been proposed in the last two decades [3,4], most of them are originally designed to handle one or fewer types of holes. For example, to repair the mesh holes generated by the Poisson surface reconstruction algorithm towards 3D printing, Sheng et al proposed an iterative postprocessing algorithm for partial differential equation (PDE) patch generation based on biharmonic-like four-order PDEs [5].…”

Amodal Completion Revisited

“…The fastgrowing neuroscientific evidence relevant to AC and related processes has been reviewed by Thielen et al (2019). As regards artificial vision, some computational models embody AC processes (Follman et al, 2018;Oliver et al, 2016;Zhu et al, 2017), while occasional references to AC are found in the broader literature on object recovery from images with missing regions, not necessarily due to occlusion (Ehsani et al, 2017;Guo et al, 2018;Han et al, 2017;Harary et al, 2014;Hueting et al, 2017;Li & Malik, 2016;Mavridis et al, 2015;Oliver et al, 2017;P erez et al, 2016;Shao et al, 2014).…”