Search

Reconstructing geometric shapes from point clouds is a common task that is often accomplished by experts manually modeling geometries in CAD-capable software. State-of-the-art workflows based on fully automatic geometry extraction are limited by point cloud density and memory constraints, and require pre- and post-processing by the user. In this work, we present a framework for interactive, user-driven, feature-assisted geometry reconstruction from arbitrarily sized point clouds. Based on seeded region-growing point cloud segmentation, the user interactively extracts planar pieces of geometry and utilizes contextual suggestions to point out plane surfaces, normal and tangential directions, and edges and corners. We implement a set of feature-assisted tools for high-precision modeling tasks in architecture and urban surveying scenarios, enabling instant-feedback interactive point cloud manipulation on large-scale data collected from real-world building interiors and facades. We evaluate our results through systematic measurement of the reconstruction accuracy, and interviews with domain experts who deploy our framework in a commercial setting and give both structured and subjective feedback., 13 pages, submitted to Computers & Graphics Journal

The Alfeld split is obtained by subdividing a simplex in R^s into s+1 sub-simplices with the barycenter as one of their vertices. On this split, we consider the space of C^1 splines of degree d = 2s−1, on which we construct a basis of simplex-splines with knots at the barycenter and the vertices of the simplex. The basis consists of two types of of simplex-splines, firstly Bernstein polynomials with domain points on the external faces of the simplex and secondly certain simplex-splines with at least one knot at the barycenter. Partition of unity, Marsdenlike identities and domain points are shown for s ≤ 20.

Despite the fast advances in high-sigma yield analysis with the help of machine learning techniques in the past decade, one of the main challenges, the curse of dimensionality, which is inevitable when dealing with modern large-scale circuits, remains unsolved. To resolve this challenge, we propose an absolute shrinkage deep kernel learning, ASDK, which automatically identifies the dominant process variation parameters in a nonlinear-correlated deep kernel and acts as a surrogate model to emulate the expensive SPICE simulation. To further improve the yield estimation efficiency, we propose a novel maximization of approximated entropy reduction for an efficient model update, which is also enhanced with parallel batch sampling for parallel computing, making it ready for practical deployment. Experiments on SRAM column circuits demonstrate the superiority of ASDK over the state-of-the-art (SOTA) approaches in terms of accuracy and efficiency with up to 10.3x speedup over SOTA methods.

Significant effort has been made to solve computationally expensive optimization problems in the past two decades, and various optimization methods incorporating surrogates into optimization have been proposed. Most research focuses on either exploiting the surrogate by defining a utility optimization problem or customizing an existing optimization method to use one or multiple approximation models. However, only a little attention has been paid to generic concepts applicable to different types of algorithms and optimization problems simultaneously. Thus this paper proposes a generalized probabilistic surrogate-assisted framework (GPSAF), applicable to a broad category of unconstrained and constrained, single- and multi-objective optimization algorithms. The idea is based on a surrogate assisting an existing optimization method. The assistance is based on two distinct phases, one facilitating exploration and another exploiting the surrogates. The exploration and exploitation of surrogates are automatically balanced by performing a probabilistic knockout tournament among different clusters of solutions. A study of multiple well-known population-based optimization algorithms is conducted with and without the proposed surrogate assistance on single- and multi-objective optimization problems with a maximum solution evaluation budget of 300 or less. The results indicate the effectiveness of applying GPSAF to an optimization algorithm and the competitiveness with other surrogate-assisted algorithms.

We describe a novel approach for exploring, visualizing and modelling of complex textile structures. We show this approach in action for the exploration of weft-knitted spacer fabrics. The modelling tools are readily available and represent a fast-growing approach in communities of generative design. The approach helps solving the problem that spacer fabrics are complex structures which are hard to imagine and to visualize. The explorations turn out insight-full and flexible and are complementary to what can be done with existing tools., Comment: 11 pages, 7 figures

Tenuous subgraph finding aims to detect a subgraph with few social interactions and weak relationships among nodes. Despite significant efforts have been made on this task, they are mostly carried out in view of graph-structured data. These methods depend on calculating the shortest path and need to enumerate all the paths between nodes, which suffer the combinatorial explosion. Moreover, they all lack the integration of neighborhood information. To this end, we propose a novel model named Graph Neural Network with Motif-aware for tenuous subgraph finding (GNNM), a neighborhood aggregation based GNNs framework which can capture the latent relationship between nodes. Specially, we design a GNN module to project nodes into a low dimensional vector combining the higher-order correlation within nodes based on a motif-aware module. And then design greedy algorithms in vector space to obtain a tenuous subgraph whose size is greater than a specified constraint. Particularly, considering existing evaluation indicators cannot capture the latent friendship between nodes, we introduce a novel Potential Friend(PF) concept to measure the tenuity of a graph from a new perspective. Experimental results on the real-world and synthetic datasets demonstrate that our proposed method GNNM outperforms existing algorithms in efficiency and subgraph quality., Comment: 15 pages, 14 figures

The modeling and control of complex physical systems are essential in real-world problems. We propose a novel framework that is generally applicable to solving PDE-constrained optimal control problems by introducing surrogate models for PDE solution operators with special regularizers. The procedure of the proposed framework is divided into two phases: solution operator learning for PDE constraints (Phase 1) and searching for optimal control (Phase 2). Once the surrogate model is trained in Phase 1, the optimal control can be inferred in Phase 2 without intensive computations. Our framework can be applied to both data-driven and data-free cases. We demonstrate the successful application of our method to various optimal control problems for different control variables with diverse PDE constraints from the Poisson equation to Burgers' equation., 15 pages, 12 figures. This paper is accepted to the Thirty-Sixth AAAI Conference on Artificial Intelligence (AAAI-22)

The construction of parametric curve and surface plays important role in computer aided geometric design (CAGD), computer aided design (CAD), and geometric modeling. In this paper, we define a new kind of blending functions associated with a real points set, called generalized toric-Bernstein (GT-Bernstein) basis functions. Then the generalized toric-Bezier (GT-B\'ezier) curves and surfaces are constructed based on the GT-Bernstein basis functions, which are the projections of the (irrational) toric varieties in fact and the generalizations of the classical rational B\'ezier curves and surfaces and toric surface patches. Furthermore, we also study the properties of the presented curves and surfaces, including the limiting properties of weights and knots. Some representative examples verify the properties and results., Comment: 28 pages, many figures

We present an observational study to compare co-located and situated real-time visualizations in basketball free-throw training. Our goal is to understand the advantages and concerns of applying immersive visualization to real-world skill-based sports training and to provide insights for designing AR sports training systems. We design both a situated 3D visualization on a head-mounted display and a 2D visualization on a co-located display to provide immediate visual feedback on a player's shot performance. Using a within-subject study design with experienced basketball shooters, we characterize user goals, report on qualitative training experiences, and compare the quantitative training results. Our results show that real-time visual feedback helps athletes refine subsequent shots. Shooters in our study achieve greater angle consistency with our visual feedback. Furthermore, AR visualization promotes an increased focus on body form in athletes. Finally, we present suggestions for the design of future sports AR studies., Comment: To appear in the 2021 ACM Conference on Human Factors in Computing Systems (CHI 2021)

Computer-aided diagnosis of lung segmentation is the fundamental requirement to diagnose lung diseases. In this paper, a two-dimensional (2D) Otsu algorithm by Darwinian particle swarm optimization (DPSO) and fractional-order Darwinian particle swarm optimization (FODPSO) is proposed to segment the pulmonary parenchyma from the lung image obtained through computed tomography (CT) scans. The proposed method extracts pulmonary parenchyma from multi-sliced CT. This is a preprocessing step to identify pulmonary diseases such as emphysema, tumor, and lung cancer. Image segmentation plays a significant role in automated pulmonary disease diagnosis. In traditional 2D Otsu, exhaustive search plays an important role in image segmentation. However, the main disadvantage of the 2D Otsu method is its complex computation and processing time. In this paper, the 2D Otsu method optimized by DPSO and FODPSO is developed to reduce complex computations and time. The efficient segmentation is very important in object classification and detection. The particle swarm optimization (PSO) method is widely used to speed up the computation and maintain the same efficiency. In the proposed algorithm, the limitation of PSO of getting trapped in local optimum solutions is overcome. The segmentation technique is assessed and equated with the traditional 2D Otsu method. The test results demonstrate that the proposed strategy gives better results. The algorithm is tested on the Lung Image Database Consortium image collections.

Optimal Bayesian design techniques provide an estimate for the best parameters of an experiment in order to maximize the value of measurements prior to the actual collection of data. In other words, these techniques explore the space of possible observations and determine an experimental setup that produces maximum information about the system parameters on average. Generally, optimal Bayesian design formulations result in multiple high-dimensional integrals that are difficult to evaluate without incurring significant computational costs as each integration point corresponds to solving a coupled system of partial differential equations. In the present work, we propose a novel approach for development of polynomial chaos expansion (PCE) surrogate model for the design utility function. In particular, we demonstrate how the orthogonality of PCE basis polynomials can be utilized in order to replace the expensive integration over the space of possible observations by direct construction of PCE approximation for the expected information gain. This novel technique enables the derivation of a reasonable quality response surface for the targeted objective function with a computational budget comparable to several single-point evaluations. Therefore, the proposed technique reduces dramatically the overall cost of optimal Bayesian experimental design. We evaluate this alternative formulation utilizing PCE on few numerical test cases with various levels of complexity to illustrate the computational advantages of the proposed approach., 30 pages, 9 figures. Published in Computer Methods in Applied Mechanics and Engineering

In order to tackle the problem of shape design and shape adjustment of complex surfaces in engineering, continuity conditions between generalized Bézier-like surfaces with multiple shape parameters are studied in this paper. Firstly, the geometric model of the generalized Bézier-like surfaces is built by blending a number of Bézier-like curves with independent shape parameters. Secondly, based on the terminal properties and linear independence of Bernstein-like basis functions, the conditions for G2 continuity between two adjacent generalized Bézier-like surfaces are derived, and then simplified by choosing appropriate shape parameters. Finally, some properties and applications of the smooth continuity between generalized Bézier-like surfaces are discussed. The modeling examples show that the proposed method is effective and easy to implement, which can greatly improve the ability to construct complex surfaces by using the generalized Bézier-like surfaces.

As a new method of representing curves, Q-Bézier curves not only exhibit the beneficial properties of Bézier curves but also allow effective shape adjustment by changing multiple shape parameters. In order to resolve the problem of not being able to construct complex curves using a single curve, we study the geometric continuity conditions for Q-Bézier curves of degree n. Following the analysis of basis functions and terminal properties of Q-Bézier curves of degree n, the continuity conditions of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$\mathrm{G}^{1}$\end{document}G1 and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$\mathrm{G}^{2}$\end{document}G2 between two adjacent Q-Bézier curves are proposed. In addition, we discuss the specific steps of smooth continuity for Q-Bézier curves and analyze the influence rules of shape parameters for Q-Bézier curves. The modeling examples show that the proposed method is effective and easy to achieve, making it useful for constructing complex curves for engineering design.

To cope with the complex embedded system design, early design space exploration (DSE) is used to make design decisions early in the design phase. For early DSE it is crucial that the running time of the exploration is as small as possible. In this paper, we describe both the porting of our scenario-based DSE to the SPARC T3-4 server and the analysis of its performance behavior.

One crucial step in virtual drug design is the identification of new lead structures with respect to a pharmacological target molecule. The search for new lead structures is often done with the help of a pharmacophore, which carries the essential structural as well as physico-chemical properties that a molecule needs to have in order to bind to the target molecule. In the absence of the target molecule, such a pharmacophore can be established by comparison of a set of active compounds. In order to identify their common features, a multiple alignment of all or most of the active compounds is necessary. Moreover, since the "outer shape" of the molecules plays a major role in the interaction between drug and target, an alignment algorithm aiming at the identification of common binding properties needs to consider the molecule's "outer shape", which can be approximated by the solvent excluded surface. In this thesis, we present a new approach to molecular surface alignment based on a discrete representation of shape as well as physico-chemical properties by points distributed on the solvent excluded surface. We propose a new method to distribute points regularly on a surface w.r.t. a smoothly varying point density given on that surface. Since the point distribution algorithm is not restricted to molecular surfaces, it might also be of interest for other applications. For the computation of pairwise surface alignments, we extend an existing point matching scheme to surface points, and we develop an efficient data structure speeding up the computation by a factor of three. Moreover, we present an approach to compute multiple alignments from pairwise alignments, which is able to handle a large number of surface points. All algorithms are evaluated on two sets of molecules: eight thermolysin inhibitors and seven HIV-1 protease inhibitors. Finally, we compare the results obtained from surface alignment with the results obtained by applying an atom alignment approach., Die Identifizierung neuer Leitstrukturen (lead structures) zur Entwicklung optimierter Wirkstoffe ist ein äußerst wichtiger Schritt in der virtuellen Wirkstoffentwicklung (virtual drug design). Die Suche nach neuen Leitstrukturen wird oft mit Hilfe eines Pharmakophor-Modells durchgeführt, welches die wichtigsten strukturellen wie auch physiko-chemischen Eigenschaften eines bindenden Moleküls in sich vereint. Ist das Zielmolekül (target) nicht bekannt, kann das Pharmakophor-Modell mit Hilfe des Vergleiches aktiver Moleküle erstellt werden. Hier ist insbesondere die gleichzeitige Überlagerung (multiple alignment) aller oder nahezu aller Moleküle notwendig. Da bei der Interaktion zweier Moleküle die "äußere Form" der Moleküle eine besondere Rolle spielt, sollte diese von jedem Überlagerungsalgorithmus, der sich mit der Identifizierung von Bindungseigenschaften befasst, berücksichtigt werden. Dabei kann die "äußere Form" durch eine bestimmte Art von molekularer Oberfläche approximiert werden, die man als solvent excluded surface bezeichnet. In dieser Arbeit stellen wir einen neuen Ansatz zur Überlagerung molekularer Oberflächen dar, der auf einer diskreten Repräsentation sowohl der Form als auch der molekularen Eigenschaften mittels Punkten beruht. Um die Punkte auf der molekularen Oberfläche möglichst regulär entsprechend einer gegebenen Punktdichte zu verteilen, entwickeln wir eine neue Methode. Diese Methode ist nicht auf Moleküloberflächen beschränkt und könnte daher auch für andere Anwendungen von Interesse sein. Basierend auf einem bekannten Point-Matching Verfahren entwickeln wir einen Point-Matching Algorithmus für Oberflächenpunkte. Dazu erarbeiten wir u.a. eine effiziente Datenstruktur, die den Algorithmus um einen Faktor von drei beschleunigt. Darüberhinaus stellen wir einen Ansatz vor, der Mehrfachüberlagerungen (multiple alignments) aus paarweisen Überlagerungen berechnet. Die Herausforderung besteht hierbei vor allem in der großen Anzahl von Punkten, die berücksichtigt werden muss. Die vorgestellten Algorithmen werden an zwei Gruppen von Molekülen evaluiert, wobei die erste Gruppe aus acht Thermolysin Inhibitoren besteht, die zweite aus sieben HIV-1 Protease Inhibitoren. Darüberhinaus vergleichen wir die Ergebnisse der Oberflächenüberlagerung mit denen einer Atommittelpunktüberlagerung.

Environmental changes, failures, collisions or even terrorist attacks can cause serious malfunctions of the delivery systems. We have presented a novel approach improving resilience of Autonomous Moving Platforms AMPs. The approach is based on multi-level state diagrams describing environmental trigger specifications, movement actions and synchronization primitives. The upper level diagrams allowed us to model advanced interactions between autonomous AMPs and detect irregularities such as deadlocks live-locks etc. The techniques were presented to verify and analyze combined AMPs' behaviors using model checking technique. The described system, Dedan verifier, is still under development. In the near future, a graphical form of verified system representation is planned., Comment: 11 pages, 5 figures

The case study analyzed in the report involves the behavioral specification and verification of a three-stage pipeline consisting of mutually concurrent modules which also compete for a shared resource. The system components are specified in terms of Concurrent State Machines (CSM) and the verification technique used is the temporal model checking in the COSMA environment., Comment: 15 pages, 11 figures, European Congress on Computational Methods in Applied Sciences and Engineering ECCOMAS 2004, Jyv\"askyl\"a, Finland, 24-28 July 2004