Variational Autoencoding of Dental Point Clouds

Interpolation and regression of data and smoothing of noisy data play a significant role in many scientific disciplines. Interpolation is an estimation of an unknown variable at output points (locations) by employing the known values at surrounding input points. Regression is an estimation of a variable at both input and output points by employing the known values at the surrounding input locations. Smoothing is an estimation of a variable at only known input points by employing the known input values. Smoothing may be necessary if the input data is noisy. It is worth noting that interpolation is different than regression and smoothing; the estimation based on interpolation passes through all the known input values. In other words, there is an exact recovery of the known values of the input points. There exist several methods for such estimations.

Point clouds consist of a discrete set of points irregularly sampled from continuous 3D objects. Most existing approaches for point cloud learning are in (semi)-supervised fashions, which nevertheless require costly human annotations. To this end, we propose a novel unsupervised learning of geometric sampling invariant representations, aiming to learn intrinsic feature representations of point clouds on graphs based on that the geometry of one object can be sampled in various patterns and densities into different forms of point clouds. In particular, we represent point clouds on graphs and exploit invariant representations at multiple hierarchies: the low-resolution invariance and original-resolution invariance. To learn invariance at a lower resolution, we subsample the input point cloud in distinct patterns, and maximize the mutual information among the subsampled variants. Further, to learn invariance at the original resolution, we increase the resolution of the subsampled point clouds to the original resolution of the input based on the learned features, and minimize the distance between the input and each of the upsampled versions. In experiments, we apply the learned representations to representative downstream tasks of point clouds, and results on point cloud classification, segmentation and upsampling demonstrate the superiority of the proposed model.

Point cloud enhancement optimization and high-fidelity texture reconstruction methods for air material via fusion of 3D scanning and neural rendering

This paper presents new techniques for vibration transmission analysis on the finite element (FE) model using Transfer Path Analysis (TPA). The techniques are 1) a contribution calculation technique for structure with manifold and continuous transfer paths, and 2) a visualization technique of contributions for efficient derivation of measures to reduce vibration. Taking advantage of the characteristics of the numerical model, the cross section whose contributions are calculated is changed sequentially between input point and output point. This contribution of each node and each DOF is converted to influence degree, which is a new indicator and expresses vibration characteristics effectively. The magnitude of the influence degree is visually expressed by deformation of the FE model to support understanding of vibration transmission characteristics. First, the authors applied the proposed techniques to the simple vehicle body FE model for the purpose of reducing vibration at the position on the panel. The parts that are responsible for vibration at the target position were shown as characteristic diagrams. On the basis of these diagrams, the authors modified the structure of the body. As a result, the panel vibration in the focused-on frequency was reduced, and the relationship was obtained by which the measures based on influence degree change the mode shape of vibration to reduce the vibration response. Next, the authors applied this approach to the W/B FE model that represents the actual vehicle body, which is complicated in the same way. As a result, the vibration at the target position was reduced. From this result, the effectiveness and usefulness of the proposed approach was demonstrated.

Stiffness characterization and mechanical manipulation of rigid or flexible micro objects using compliant devices are investigated in this paper. Compliant devices are themselves elastically deformable objects. By using the material properties, geometry, and measured deformations of the compliant devices, the force and deflection of the manipulated object are computed using geometrically nonlinear mechanics formulations. Thus, the stiffness of any given object is characterized. Since the input and output points might be occluded, we have developed an alternate approach where displacement measurement at other points can be used to estimate the forces at input and output points. This new technique also helps reduce the effect of errors in optical capturing of deformations in the computation of forces. By virtue of this vision-based force measurement, we can apply controlled force on the objects, which is a very useful feature in micro assembly and manipulation where excessive force might damage the object. Additionally, we demonstrate manipulation with compliant devices utilizing vision-guided position and force control. Illustrative examples at micro and macro size scales are included.

Adversary and invisibility are two fundamental but conflict characters of adversarial perturbations. Previous adversarial attacks on 3D point cloud recognition have often been criticized for their noticeable point outliers, since they just involve an “implicit constrain” like global distance loss in the time-consuming optimization to limit the generated noise. While point cloud is a highly structured data format, it is hard to constrain its perturbation with a simple loss or metric properly. In this paper, we propose a novel Point-Cloud Sensitivity Map to boost both the efficiency and imperceptibility of point perturbations. This map reveals the vulnerability of point cloud recognition models when encountering shape-invariant adversarial noises. These noises are designed along the shape surface with an “explicit constrain” instead of extra distance loss. Specifically, we first apply a reversible coordinate transformation on each point of the point cloud input, to reduce one degree of point freedom and limit its movement on the tangent plane. Then we calculate the best attacking direction with the gradients of the transformed point cloud obtained on the white-box model. Finally we assign each point with a non-negative score to construct the sensitivity map, which benefits both white-box adversarial invisibility and black-box query-efficiency extended in our work. Extensive evaluations prove that our method can achieve the superior performance on various point cloud recognition models, with its satisfying adversarial imperceptibility and strong resistance to different point cloud defense settings. Our code is available at: https://github.com/shikiw/SI-Adv.

Introduction and aimsIntraoral scanning (IOS) captures real-time surface morphology of teeth and is widely used in clinical dentistry. However, due to the complex intraoral environment, data loss during scanning is common, leading to incomplete three-dimensional (3D) point clouds. This study aimed to develop and evaluate a deep learning–based method to automatically restore missing regions in intraoral 3D point clouds, thereby improving the accuracy and efficiency of digital orthodontic workflows.MethodsA Point Fractal Network architecture was adopted to reconstruct incomplete IOS data. A dataset comprising 314 IOS scans and 4162 individual teeth was used for training and validation. Missing data were simulated by removing random portions of point clouds (5%, 10%, 15%, and 20%). Model performance was assessed using Chamfer distance (CD) to measure the accuracy of point cloud completion across different levels of data loss.ResultsThe proposed method achieves robust performance, maintaining average CD values below 0.01 across most levels of simulated data loss. Visual comparisons confirmed high geometric fidelity between the completed and original point clouds. Furthermore, the model demonstrated efficient processing, completing each point cloud in approximately 0.5 seconds, enabling near real-time restoration during clinical scanning.ConclusionThe deep learning–based model accurately restores missing IOS data, improving the precision and efficiency of digital dental workflows. Its speed and accuracy support real-time clinical applications and reduce reliance on manual corrections.Clinical RelevanceThis method improves clinical efficiency, reduces chairside time, and enhances both patient comfort and treatment acceptance. In addition, it minimises human error and increases the precision of dental restorations. As digital dentistry continues to evolve, this approach holds great potential for improving the accuracy and efficiency of dental treatments, paving the way for broader artificial intelligence integration in clinical practice.

Chapter 10 - DDC Introduction to Hardware and Software

Point cloud completion aims to reconstruct the complete 3D shape from incomplete point clouds, and it is crucial for tasks such as 3D object detection and segmentation. Despite the continuous advances in point cloud analysis techniques, feature extraction methods are still confronted with apparent limitations. The sparse sampling of point clouds, used as inputs in most methods, often results in a certain loss of global structure information. Meanwhile, traditional local feature extraction methods usually struggle to capture the intricate geometric details. To overcome these drawbacks, we introduce PointCFormer, a transformer framework optimized for robust global retention and precise local detail capture in point cloud completion. This framework embraces several key advantages. First, we propose a relation-based local feature extraction method to perceive local delicate geometry characteristics. This approach establishes a fine-grained relationship metric between the target point and its k-nearest neighbors, quantifying each neighboring point's contribution to the target point's local features. Secondly, we introduce a progressive feature extractor that integrates our local feature perception method with self-attention. Starting with a denser sampling of points as input, it iteratively queries long-distance global dependencies and local neighborhood relationships. This extractor maintains enhanced global structure and refined local details, without generating substantial computational overhead. Additionally, we develop a correction module after generating point proxies in the latent space to reintroduce denser information from the input points, enhancing the representation capability of the point proxies. PointCFormer demonstrates state-of-the-art performance on several widely used benchmarks.

Traditional Convolutional Neural Networks (CNN) are limited to extract informative local features of point clouds due to the fixed geometric structures in convolution kernel against irregular and unstructured point clouds. It usually requires data transformation such as voxelization or projection, inducing a possible loss of information. Instead of fitting the input points to the kernel by regularization, we choose to fit the kernel to input points to conduct convolution. In this paper, we present a new method to define and compute convolution directly on 3D point clouds by Adaptive Surface Fitting Convolution (ASFConv). The key idea is to utilize a set of kernel points distributed on the tangent plane and project them back to point cloud surface. After adapting to the distribution of input points, ASFConv kernel can better capture local neighborhood geometry and benefit the feature extraction. In the experiments, we evaluate our network on two public datasets: ModelNet40 and ShapeNet for classification and segmentation. The experimental results show that our method obtain competitive performances compared to the state-of-the-art.

Point cloud compression (PCC) is crucial for efficient and flexible storage as well as feasible transmission of point clouds in practice. For geometry compression, one popular approach is the octree-based solution. The intra prediction mechanism utilizes the spatial correlation in the static point cloud to predict the occupancy bit of the octree node for entropy coding, reducing the spatial redundancy. In this study, two local geometry-based prediction methods are proposed following statistical and theoretical analyses: binary prediction, which outputs the binary state (i.e., occupied or unoccupied), and ternary prediction, which provides a third option other than occupied or unoccupied (i.e., not predicted). In comparison to the state-of-the-art, the proposed binary prediction offers the Bjontegaard delta rate (BD-rate) of −0.8% for lossy compression and the bits per input point (bpip) of 100.09% for lossless compression in average, respectively. The binary prediction reduces the computational complexity in terms of more than 20% decrease in decoding time. In particular, it also provides noticeable reduction of the memory usage during entropy coding. The proposed ternary prediction provides −1.2% BD-rate for lossy compression and 97.19% bpip for lossless compression in average, respectively, in comparison to the state-of-the-art. While achieving performance gain, it is considerably more computational efficient by saving about 18% decoding time. Due to these advantages, part of the proposed ternary prediction has been adopted by the ongoing MPEG standard of geometry-based point cloud compression (G-PCC).

We consider the problem of locating input and output (I/O) points of each department for a given layout. The objective of the problem is to minimise the total distance of material flows between the I/O points. Here, distances between the I/O points are computed as the lengths of the shortest path (not the rectilinear distances) between the I/O points. We developed a procedure to eliminate dominated candidate positions of I/O points that do not need to be considered. With this procedure, a large number of dominated candidate positions can be eliminated. A linear programming (LP) model for minimising the total rectilinear distance of flows is used to obtain a lower bound. Using the elimination procedure and the LP model, a branch and bound algorithm is developed to find an optimal location of the I/O points. Results from computational experiments show that the suggested algorithm finds optimal solutions in a very short time even for large-sized problems.

Chapter Thirteen - Inferring Transforms: May-June 1999

ChatIOS: Improving automatic 3-dimensional tooth segmentation via GPT-4V and multimodal pre-training.

This paper presents a general Surface Reconstruction framework which encapsulates the uncertainty of the sam-pled data, making no assumption on the shape of the surface to be reconstructed. Starting from the input points(either points clouds or multiple range images), an Estimated Existence Function (EEF) is built which modelsthe space in which the desired surface could exist and, by the extraction of EEF critical points, the surface isreconstructed. The nal goal is the development of a generic framework able to adapt the result to different kindsof additional information coming from multiple sensors. Categories and Subject Descriptors (according to ACM CCS) : I.3.3 [Computer Graphics]: Shape Modeling, Uncer-tain data, Multi-sensor Data Fusion 1. Introduction 3D scanning devices are becoming more and more availableand affordable. Thanks to modern acquisition technologies,heterogeneous data can be acquired from multiple acquisi-tion sensors, which often incorporate information about un-certainty of the data sampling process. Surface reconstruc-tion techniques designed over a specic sensor often takeinto account uncertainty during the reconstruction process,but they are limited to work with a single device. On thecontrary, general techniques that can process data comingfrom different sensors usually disregard much part of sensor-specic information, and seldom take into account uncer-tainty.The basic concept of our approach is