CEIG19

Permanent URI for this collection

https://diglib.eg.org/handle/10.2312/2632793

Browse

Now showing 1 - 10 of 10

3D Environment Understanding in Real-time Using Input CAD Models for AR Applications
(The Eurographics Association, 2019) Rodríguez, David Jurado; Rodríguez, Juan Manuel Jurado; Alvarado, Lidia Ortega; Higueruela, Francisco Ramón Feito; Casas, Dan and Jarabo, Adrián
The automatic recognition of real environments has become a relevant issue for multiple purposes in computer graphics, computer vision and artificial intelligence. In this work, we focus on environment understanding according to input CAD models for an Augmented Reality (AR) application. We provide a novel solution for the management of building infrastructures in indoor spaces. The use case has been the University of Jaén to visualize correctly their service facilites from AR. To this end, firstly, the CAD models (2D) have been segmented in order to simplify its geometry. As a result, an efficient data structure has been created for real-time alignement to scanned data. Secondly, we have developed a mobile application based on ARCore library to capture and generate 3D planes of the user environment. Finally, we have carried out a method to align automatically the virtual elements such as walls, doors and grounds to the real world. The main objective of this research is to calculate the needed geometric transformations of virtual elements and thus, to achieve a correct overlappping with the real world by understanding their physical and spatial constraints in real time.
A Level-of-Detail Technique for Urban Physics Calculations in Large Urban Environments
(The Eurographics Association, 2019) Novoa, David Muñoz; Besuievsky, Gonzalo; Patow, Gustavo; Casas, Dan and Jarabo, Adrián
In many applications, such as urban physics simulations or the study of the solar impact effects at different scales, complex 3D city models are required to evaluate physical values. In this paper we present a new technique which, through the use of an electrical analogy and the calculation of sky view factors and form factors, allows to simulate and study the thermal behaviour of an urban environment, taking into account the solar and sky radiation, the air and sky temperatures, and even the thermal interaction between nearby buildings. We also show that it is possible, from a 3D recreation of a large urban environment, to simulate the heat exchanges that take place between the buildings of a city and its immediate surroundings. In the same way, taking into account the terrestrial zone, the altitude and the type of climate with which the simulations are carried out, it is possible to compare the thermal behaviour of a large urban environment according to the chosen conditions.
Modal Locomotion of C. elegans
(The Eurographics Association, 2019) Mujika, Andoni; Merino, Sara; Leškovský, Peter; Epelde, Gorka; Oyarzun, David; Otaduy, Miguel Angel; Casas, Dan and Jarabo, Adrián
Caenorhabditis elegans (C. elegans) is a roundworm that, thanks to its combination of biological simplicity and behavioral richness, offers an excellent opportunity for initial experimentation of many human diseases. In this work, we introduce a locomotion model for C. elegans, which can enable in-silico validation of behavioral experiments prior to physical experimentation with actual C. elegans specimens. Our model enables interactive simulation of self-propelling C. elegans, using as sole input biologically inspired muscle forces and frictional contact. The key to our model is a simple locomotion control strategy that activates selected natural vibration modes of the worm. We perform an offline analysis of the natural vibration modes, select those that best match the deformation of the worm during locomotion, and design force profiles that activate these vibration modes in a coordinated manner. Together with force compensation for momentum conservation and an anisotropic friction model, we achieve locomotions that match qualitatively those of real-world worms. Our approach is general, and could be extended to the locomotion of other types of animals or characters.
Perfect Spatial Hashing for Point-cloud-to-mesh Registration
(The Eurographics Association, 2019) Mejia-Parra, Daniel; Lalinde-Pulido, Juan; Sánchez, Jairo R.; Ruiz-Salguero, Oscar; Posada, Jorge; Casas, Dan and Jarabo, Adrián
Point-cloud-to-mesh registration estimates a rigid transformation that minimizes the distance between a point sample of a surface and a reference mesh of such a surface, both lying in different coordinate systems. Point-cloud-to-mesh-registration is an ubiquitous problem in medical imaging, CAD CAM CAE, reverse engineering, virtual reality and many other disciplines. Common registration methods include Iterative Closest Point (ICP), RANdom SAmple Consensus (RANSAC) and Normal Distribution Transform (NDT). These methods require to repeatedly estimate the distance between a point cloud and a mesh, which becomes computationally expensive as the point set sizes increase. To overcome this problem, this article presents the implementation of a Perfect Spatial Hashing for point-cloud-to-mesh registration. The complexity of the registration algorithm using Perfect Spatial Hashing is O(NYxn) (NY : point cloud size, n: number of max. ICP iterations), compared to standard octrees and kd-trees (time complexity O(NY log(NT)xn), NT : reference mesh size). The cost of pre-processing is O(NT +(N3H )2) (N3H : Hash table size). The test results show convergence of the algorithm (error below 7e-05) for massive point clouds / reference meshes (NY = 50k and NT = 28055k, respectively). Future work includes GPU implementation of the algorithm for fast registration of massive point clouds.
Procedural Location of Roads Using Desire Paths
(The Eurographics Association, 2019) Real, Pablo; Martínez-Gil, Francisco; Martínez-Durá, Rafael J.; García-Fernández, Ignacio; Casas, Dan and Jarabo, Adrián
Procedural modelling of realistic environments that include elements derived from human activity can largely reduce production cost in animation, video-games and feature films. We address the problem of placing roads and other human-made elements, such as buildings, in a way that is consistent with the scene relief. Our approach is based on the calculation of so called desire paths, by means of the generation of many optimal paths according to different cost or distance functions. Using this idea, we rely on the information of the terrain properties to emulate the exploration of the scenario by a large number of pedestrians. From the routes generated by the pedestrians, we define walkability and habitability maps on the scene that can be later used to locate roads or buildings.
Sensitivity Analysis in Shape Optimization using Voxel Density Penalization
(The Eurographics Association, 2019) Montoya-Zapata, Diego; Acosta, Diego A.; Moreno, Aitor; Posada, Jorge; Ruiz-Salguero, Oscar; Casas, Dan and Jarabo, Adrián
Shape optimization in the context of technical design is the process by which mechanical demands (e.g. loads, stresses) govern a sequence of piece instances, which satisfy the demands, while at the same time evolving towards more attractive geometric features (e.g. lighter, cheaper, etc.). The SIMP (Solid Isotropic Material with Penalization) strategy seeks a redistribution of local densities of a part in order to stand stress / strain demands. Neighborhoods (e.g. voxels) whose density drifts to lower values are considered superfluous and removed, leading to an optimization of the part shape. This manuscript presents a study on how the parameters governing the voxel pruning affect the convergence speed and performance of the attained shape. A stronger penalization factor establishes the criteria by which thin voxels are considered void. In addition, the filter discourages punctured, chessboard pattern regions. The SIMP algorithm produces a forecasted density map on the whole piece voxels. A post-processing is applied to effectively eliminate voxels with low density, to obtain the effective shape. In the literature, mechanical performance finite element analyses are conducted on the full voxel set with diluted densities by linearly weakening each voxel resistance according to its diluted density. Numerical tests show that this approach predicts a more favorable mechanical performance as compared with the one obtained with the shape which actually lacks the voxels with low density. This voxel density - based optimization is particularly convenient for additive manufacturing, as shown with the piece actually produced in this work. Future endeavors include different evolution processes, albeit based on variable density voxel sets, and mechanical tests conducted on the actual sample produced by additive manufacture.
View-dependent Hierarchical Rendering of Massive Point Clouds through Textured Splats
(The Eurographics Association, 2019) Comino Trinidad, Marc; Calaf, Antonio Chica; Gran, Carlos Andújar; Casas, Dan and Jarabo, Adrián
Nowadays, there are multiple available range scanning technologies which can capture extremely detailed models of realworld surfaces. The result of such process is usually a set of point clouds which can contain billions of points. While these point clouds can be used and processed offline for a variety of purposes (such as surface reconstruction and offline rendering) it is unfeasible to interactively visualize the raw point data. The most common approach is to use a hierarchical representation to render varying-size oriented splats, but this method also has its limitations as usually a single color is encoded for each point sample. Some authors have proposed the use of color-textured splats, but these either have been designed for offline rendering or do not address the efficient encoding of image datasets into textures. In this work, we propose extending point clouds by encoding their color information into textures and using a pruning and scaling rendering algorithm to achieve interactive rendering. Our approach can be combined with hierarchical point-based representations to allow for real-time rendering of massive point clouds in commodity hardware.
A Virtual Reality Front-end for City Modeling
(The Eurographics Association, 2019) Rando, Eduard; Andujar, Carlos; Patow, Gustavo A.; Casas, Dan and Jarabo, Adrián
Current tools for city modeling, including those supporting procedural techniques, have a steep learning curve and require substantial user input and/or skills to create realistic 3D models of cities. In this paper we propose a VR tool for the fast and intuitive creation of 3D models of cities through their main elements (buildings, blocks and streets). The key ingredients of our approach are: (a) intuitive creation of mass volume models for buildings, whose facades can be refined later on through procedural rules, (b) the ability to use arbitrary urban layouts, either created from scratch within the tool or imported from public map services, (c) algorithms to replicate and transfer user-generated blocks to arbitrary block shapes, so that a few template blocks suffice to cover the whole urban layout. The major benefit of our approach is that city design and inspection tasks are done simultaneously in a completely immersive environment.
A Virtual Reality Front-End for Earthquake Simulation
(The Eurographics Association, 2019) Fita, Josep Lluis; Besuievsky, Gonzalo; Patow, Gustavo A.; Casas, Dan and Jarabo, Adrián
Virtual Reality has traditionally been used in Cultural Heritage for giving to the final user an immersive experience over recreated scenarios, which usually have been designed and focused on static environment recreation. In spite of its importance for cultural heritage, we have found a lack of virtual reality applications to recreate structural and seismic simulations on historical buildings. In this paper we describe a low-cost virtual reality solution, affordable for all kinds of users that own a smart-phone. Through our application, the users can have an immersive experience that combines the ancient building recreation, its structural simulation and the natural phenomena simulation like earthquakes.
A Voxel-based Deep Learning Approach for Point Cloud Semantic Segmentation
(The Eurographics Association, 2019) Díaz-Medina, Miguel; Fuertes-García, José Manuel; Ogayar-Anguita, Carlos Javier; Lucena, Manuel; Casas, Dan and Jarabo, Adrián
Semantic segmentation has been a research topic in computer vision for decades. This task has become a crucial challenge nowadays due to emergence of new technologies such as autonomous driving. Nonetheless, most existing segmentation methods are not designed for handling the unstructured and irregular nature of 3D point clouds. We propose a voxel-based technique for point cloud data semantic segmentation of 3D point clouds using 3D convolutional neural networks. It uses local voxelizations for learning spatial patterns, and also corrects the imbalance of the data, something very problematic with 3D datasets.

Browse

Browsing CEIG19 by Subject "Computing methodologies"

Results Per Page

Sort Options