Gaussian splats provide a fast, high-fidelity representation for 3D objects but are often constructed from incomplete input data in practice, leaving missing regions. Existing completion methods either do not preserve the original splats or require scarcely available 3D training data. We propose GSComplete, which combines 3D generation based on Score Distillation Sampling with a novel preservation loss that encourages the original splats to be preserved where they should be visible. This effectively completes the Gaussian splat object using only 2D diffusion priors while fully preserving existing splats and generating new splats only in missing regions, without occluding the input. To evaluate our approach, we introduce a new dataset of partial Gaussian splat objects and show that GSComplete achieves significantly more accurate preservation of the input than existing methods with comparable plausibility of the completed result. Our code and dataset will be made available upon acceptance.

Point clouds, usually obtained through scanning or various image processing, are commonly affected by noise and outliers. Such artifacts compromise data quality as they significantly distort subsequent processes, such as normal estimation and surface reconstruction. In this work, we introduce a proximity-based outlier removal method for point clouds. We improve on statistical methods based on neighboring graphs by using a parameter-free proximity graph—the spheres-of-influence (SIG), thus requiring fewer parameters compared to classical methods and obtaining better results. Moreover, the simplicity of our method allows it to become an easy replacement for existing statistical methods.

As the accessibility of high-quality 3D scans increases, processing the scanned data becomes more challenging. 3D scanners obtain very large, unstructured sets of points, so called point clouds. To be able to use the data in a meaningful way it is necessary to reconstruct the surface of the scanned object from the point cloud, resulting in a 3D model. This is called the problem of 3D surface reconstruction. Processing very large point clouds (in a reasonable time) is necessary in order to keep up with ever increasing scanning technology. In this thesis, we construct, implement and evaluate a distributed surface reconstruction algorithm called DistributedBallFilter. It is a distributed-memory parallel version of the recently developed BallFilter algorithm [Ohr22]. Firstly, the input point cloud is subdivided into chunks called tiles using a 3D grid. To ensure the correctness of the results, tiles are slightly overlapping on their borders. After splitting the input, each tile can be processed independently from each other. The tiles are assigned and distributed to a number of p processes. The assignment of tiles to processes is calculated using longest-processing-time-first list scheduling. Then all processes reconstruct the 3D surface of all their assigned tiles in parallel. After all tiles are processed, the result is merged back together into a single 3D model, containing the reconstructed surface of the entire input point cloud. The asymptotic run time complexity is O(n log n) in the worst case (same as BallFilter) and O(n + n log n p ) in the best case, depending on the distribution of points within the input data. Furthermore, we implemented the algorithm in C++. The input splitting is run on a GPU using CUDA and discussed thoroughly in its dedicated paper [Bru22]. For each tile a single file is output, which is communicated to each process via a distributed file system. The MPI standard is used for sending all local results to a single process, which is also responsible for merging and outputting the final 3D model. Finally we executed the algorithm on the VSC3+ cluster, a high-performance cluster based in Vienna. It was run against several test data sets. We visualize the results and analysed the behavior of the running time when scaling the number of processes as well as the input size. In our tests, DistributedBallFilter managed to be up to around five times faster than BallFilter depending on the number of nodes used and the input size. The largest observed speedup was by a factor of 5.89 compared to BallFilter.

Determining connectivity in unstructured point clouds is a long-standing problem that is still not addressed satisfactorily. In this poster, we propose an extension to the proximity graph introduced in [MOW22] to three-dimensional models. We use the spheres-of-influence (SIG) proximity graph restricted to the 3D Delaunay graph to compute connectivity between points. Our approach shows a better encoding of the connectivity in relation to the ground truth than the k-nearest neighborhood (kNN) for a wide range of k values, and additionally, it is parameter-free. Our result for this fundamental task offers potential for many applications relying on kNN, e.g., improvements in normal estimation, surface reconstruction, motion planning, simulations, and many more.

We introduce a Delaunay-based algorithm for reconstructing the underlying surface of a given set of unstructured points in 3D. The implementation is very simple, and it is designed to work in a parameter-free manner. The solution builds upon the fact that in the continuous case, a closed surface separates the set of maximal empty balls (medial balls) into an interior and exterior. Based on discrete input samples, our reconstructed surface consists of the interface between Voronoi balls, which approximate the interior and exterior medial balls. An initial set of Voronoi balls is iteratively processed, merging Voronoi-ball pairs if they fulfil an overlapping error criterion. Our complete open-source reconstruction pipeline performs up to two quick linear-time passes on the Delaunay complex to output the surface, making it an order of magnitude faster than the state of the art while being competitive in memory usage and often superior in quality. We propose two variants (local and global), which are carefully designed to target two different reconstruction scenarios for watertight surfaces from accurate or noisy samples, as well as real-world scanned data sets, exhibiting noise, outliers, and large areas of missing data. The results of the global variant are, by definition, watertight, suitable for numerical analysis and various applications (e.g., 3D printing). Compared to classical Delaunay-based reconstruction techniques, our method is highly stable and robust to noise and outliers, evidenced via various experiments, including on real-world data with challenges such as scan shadows, outliers, and noise, even without additional preprocessing.

Reconstructing 2D curves from sample points has long been a critical challenge in computer graphics, finding essential applications in vector graphics. The design and editing of curves on surfaces has only recently begun to receive attention, primarily relying on human assistance, and where not, limited by very strict sampling conditions. In this work, we formally improve on the state-of-the-art requirements and introduce an innovative algorithm capable of reconstructing closed curves directly on surfaces from a given sparse set of sample points. We extend and adapt a state-of-the-art planar curve reconstruction method to the realm of surfaces while dealing with the challenges arising from working on non-Euclidean domains. We demonstrate the robustness of our method by reconstructing multiple curves on various surface meshes. We explore novel potential applications of our approach, allowing for automated reconstruction of curves on Riemannian manifolds.

We introduce the Stability-Incorporated Neighborhood Graph (SING), a novel density-aware structure designed to capture the intrinsic geometric properties of a point set. We improve upon the spheres-of-influence graph by incorporating additional features to offer more flexibility and control in encoding proximity information and capturing local density variations. Through persistence analysis on our proximity graph, we propose a new clustering technique and explore additional variants incorporating extra features for the proximity criterion. Alongside the detailed analysis and comparison to evaluate its performance on various datasets, our experiments demonstrate that the proposed method can effectively extract meaningful clusters from diverse datasets with variations in density and correlation. Our application scenarios underscore the advantages of the proposed graph over classical neighborhood graphs, particularly in terms of parameter tuning.

Determining connectivity between points and reconstructing their shape boundaries are long‐standing problems in computer graphics. One possible approach to solve these problems is to use a proximity graph. We propose a new proximity graph computed by intersecting the to‐date rarely used proximity‐based graph called spheres‐of‐influence graph (SIG) with the Delaunay triangulation (DT). We prove that the resulting graph, which we name SIGDT, contains the piece‐wise linear reconstruction for a set of unstructured points in the plane for a sampling condition superseding current bounds and capturing well practical point sets' properties. As an application, we apply a dual of boundary adjustment steps from the C onnect 2D algorithm to remove the redundant edges. We show that the resulting algorithm SIG‐C onnect 2D yields the best reconstruction accuracy compared to state‐of‐the‐art algorithms from a recent comprehensive benchmark, and the method offers the potential for further improvements, e.g., for surface reconstruction.

Large-scale unstructured point cloud scenes can be quickly visualized without prior reconstruction by utilizing levels-of-detail structures to load an appropriate subset from out-of-core storage for rendering the current view. However, as soon as we need structures within the point cloud, e.g., for interactions between objects, the construction of state-of-the-art data structures requires O(NlogN) time for N points, which is not feasible in real time for millions of points that are possibly updated in each frame. Therefore, we propose to use a surface representation structure which trades off the (here negligible) disadvantage of single-frame use for both output-dominated and near-linear construction time in practice, exploiting the inherent 2D property of sampled surfaces in 3D. This structure tightly encompasses the assumed surface of unstructured points in a set of bounding depth intervals for each cell of a discrete 2D grid. The sorted depth samples in the structure permit fast surface queries, and on top of that an occlusion graph for the scene comes almost for free. This graph enables novel real-time user operations such as revealing partially occluded objects, or scrolling through layers of occluding objects, e.g., walls in a building. As an example application we showcase a 3D scene exploration framework that enables fast, more sophisticated interactions with point clouds rendered in real time.

Curve reconstruction from unstructured points in a plane is a fundamental problem with many applications that has generated research interest for decades. Involved aspects like handling open, sharp, multiple and non-manifold outlines, run-time and provability as well as potential extension to 3D for surface reconstruction have led to many different algorithms. We survey the literature on 2D curve reconstruction and then present an open-sourced benchmark for the experimental study. Our unprecedented evaluation on a selected set of planar curve reconstruction algorithms aims to give an overview of both quantitative analysis and qualitative aspects for helping users to select the right algorithm for specific problems in the field. Our benchmark framework is available online to permit reproducing the results, and easy integration of new algorithms.

The design of functional seating furniture is a complicated process which often requires extensive manual design effort and empirical evaluation. We propose a computational design framework for pose-driven automated generation of body-supports which are optimized for comfort of sitting. Given a human body in a specified pose as input, our method computes an approximate pressure distribution that also takes frictional forces and body torques into consideration which serves as an objective measure of comfort. Utilizing this information to find out where the body needs to be supported in order to maintain comfort of sitting, our algorithm can create a supporting mesh suited for a person in that specific pose. This is done in an automated fitting process, using a template model capable of supporting a large variety of sitting poses. The results can be used directly or can be considered as a starting point for further interactive design.

A key step in any scanning-based asset creation workflow is to convert unordered point clouds to a surface. Classical methods (e.g., Poisson reconstruction) start to degrade in the presence of noisy and partial scans. Hence, deep learning based methods have recently been proposed to produce complete surfaces, even from partial scans. However, such data-driven methods struggle to generalize to new shapes with large geometric and topological variations. We present Points2Surf, a novel patch-based learning framework that produces accurate surfaces directly from raw scans without normals. Learning a prior over a combination of detailed local patches and coarse global information improves generalization performance and reconstruction accuracy. O5ur extensive comparison on both synthetic and real data demonstrates a clear advantage of our method over state-of-the-art alternatives on previously unseen classes (on average, Points2Surf brings down reconstruction error by 30% over SPR and by 270%+ over deep learning based SotA methods) at the cost of longer computation times and a slight increase in small-scale topological noise in some cases. Our source code, pre-trained model, and dataset are available at: https://github.com/ErlerPhilipp/points2surf .

We propose an efficient out‐of‐core octree generation method for arbitrarily large point clouds. It utilizes a hierarchical counting sort to quickly split the point cloud into small chunks, which are then processed in parallel. Levels of detail are generated by subsampling the full data set bottom up using one of multiple exchangeable sampling strategies. We introduce a fast hierarchical approximate blue‐noise strategy and compare it to a uniform random sampling strategy. The throughput, including out‐of‐core access to disk, generating the octree, and writing the final result to disk, is about an order of magnitude faster than the state of the art, and reaches up to around 6 million points per second for the blue‐noise approach and up to around 9 million points per second for the uniform random approach on modern SSDs.

We propose a parameter-free method to recover manifold connectivity in unstructured 2D point clouds with high noise in terms of the local feature size. This enables us to capture the features which emerge out of the noise. To achieve this, we extend the reconstruction algorithm HNN-Crust, which connects samples to two (noise-free) neighbours and has been proven to output a manifold for a relaxed sampling condition. Applying this condition to noisy samples by projecting their k-nearest neighbourhoods onto local circular fits leads to multiple candidate neighbour pairs and thus makes connecting them consistently an NP-hard problem. To solve this efficiently, we design an algorithm that searches that solution space iteratively on different scales of k. It achieves linear time complexity in terms of point count plus quadratic time in the size of noise clusters. Our algorithm FitConnect extends HNN-Crust seamlessly to connect both samples with and without noise, performs as local as the recovered features and can output multiple open or closed piecewise curves. Incidentally, our method simplifies the output geometry by eliminating all but a representative point from noisy clusters. Since local neighbourhood fits overlap consistently, the resulting connectivity represents an ordering of the samples along a manifold. This permits us to simply blend the local fits for denoising with the locally estimated noise extent. Aside from applications like reconstructing silhouettes of noisy sensed data, this lays important groundwork to improve surface reconstruction in 3D. Our open-source algorithm is available online.

We reconstruct a closed denoised curve from an unstructured and highly noisy 2D point cloud. Our proposed method uses a two- pass approach: Previously recovered manifold connectivity is used for ordering noisy samples along this manifold and express these as residuals in order to enable parametric denoising. This separates recovering low-frequency features from denoising high frequencies, which avoids over-smoothing. The noise probability density functions (PDFs) at samples are either taken from sensor noise models or from estimates of the connectivity recovered in the first pass. The output curve balances the signed distances (inside/outside) to the samples. Additionally, the angles between edges of the polygon representing the connectivity become minimized in the least-square sense. The movement of the polygon's vertices is restricted to their noise extent, i.e., a cut-off distance corresponding to a maximum variance of the PDFs. We approximate the resulting optimization model, which consists of higher-order functions, by a linear model with good correspondence. Our algorithm is parameter-free and operates fast on the local neighborhoods determined by the connectivity. We augment a least-squares solver constrained by a linear system to also handle bounds. This enables us to guarantee stochastic error bounds for sampled curves corrupted by noise, e.g., silhouettes from sensed data, and we improve on the reconstruction error from ground truth. Open source to reproduce figures and tables in this paper is available at: this https URL

We reconstruct a closed denoised curve from an unstructured and highly noisy 2D point cloud. Our proposed method uses a two-pass approach: Previously recovered manifold connectivity is used for ordering noisy samples along this manifold and express these as residuals in order to enable parametric denoising. This separates recovering low-frequency features from denoising high frequencies, which avoids over-smoothing. The noise probability density functions (PDFs) at samples are either taken from sensor noise models or from estimates of the connectivity recovered in the first pass. The output curve balances the signed distances (inside/outside) to the samples. Additionally, the angles between edges of the polygon representing the connectivity become minimized in the least-square sense. The movement of the polygon’s vertices is restricted to their noise extent, i.e., a cut-off distance corresponding to a maximum variance of the PDFs. We approximate the resulting optimization model, which consists of higher-order functions, by a linear model with good correspondence. Our algorithm is parameter-free and operates fast on the local neighborhoods determined by the connectivity. This enables us to guarantee stochastic error bounds for sampled curves corrupted by noise, e.g., silhouettes from sensed data, and we improve on the reconstruction error from ground truth. Source code is available online. An extended version is available at: https://arxiv.org/abs/1808.07778 CCS Concepts •Computing methodologies → Shape modeling; Point-based models;

Surface selection operations by a user are fundamental for many applications and a standard tool in mesh editing software. Unfortunately, defining a selection is only straightforward if the region is visible and on a convex model. Concave surfaces can exhibit self-occlusions, which require using multiple camera positions to obtain unobstructed views. The process thus becomes iterative and cumbersome. Our novel approach enables selections to lie under occlusions and even on the backside of objects and for arbitrary depth complexity at interactive rates. We rely on a user-drawn curve in screen space, which is projected onto the mesh and analyzed with respect to visibility to guarantee a continuous path on the surface. Our occlusion-aware surface-processing method enables a number of applications in an easy way. As examples, we show continuous painting on the surface, selecting regions for texturing, creating illustrative cutaways from nested models and animate them.