Search

Identifying conserved (similar) three-dimensional patterns among a set of proteins can be helpful for the rational design of polypharmacological drugs. Some available tools allow this identification from a limited perspective, only considering the available information, such as known binding sites or previously annotated structural motifs. Thus, these approaches do not look for similarities among all putative orthosteric and or allosteric bindings sites between protein structures. To overcome this tech-weakness Geomfinder was developed, an algorithm for the estimation of similarities between all pairs of three-dimensional amino acids patterns detected in any two given protein structures, which works without information about their known patterns. Even though Geomfinder is a functional alternative to compare small structural proteins, it is computationally unfeasible for the case of large protein processing and the algorithm needs to improve its performance. This work presents several parallel versions of the Geomfinder to exploit SMPs, distributed memory systems, hybrid version of SMP and distributed memory systems, and GPU based systems. Results show significant improvements in performance as compared to the original version and achieve up to 24.5x speedup when analyzing proteins of average size and up to 95.4x in larger proteins, This research was funded by DICYT-USACH grant 5392102RP-AC and the Fondo Nacional de Desarrollo Científico y Tecnológico (FONDECYT) grants 1191133 and 1170662 and from Spanish Government (projects SEV2015-0493 and TIN2015-65316-P, grant BES-2016-078046), and from Generalitat de Catalunya (contracts 2017-SGR1414 and 2017-SGR-1328). The web-server is hosted in the cluster obtained with the grant CONICYT-FONDEQUIPEQM160063., Peer Reviewed, Postprint (published version)

This document contains the description and development of a 3D rendering engine written with C++ and OpenGL. The main objective of this project was to implement advanced rendering techniques to have a performant and visually accurate renderer. The result of this project is a 3D rendering engine with 3 rendering pipelines to choose from. All of them are capable of rendering PBR assets. The application built to test the features resembles an engine editor.

Wildfires can have a big impact on our environment and nature. That is why it would be interesting to meet their behaviour. Our project is willing to create a 3D reconstruction of fire smoke plumes using a small number of images. It is done using camera parametrization to know the position, shape and scale of the object creating a Shape from Silhouette (SfS). The entry will be two images of the smoke taken at 90 degrees between them and facing the same point. The output is going to be used to initialize a convolutional neural network that will transform the object into a density and smoothen the borders. It has been observed that using a SfS as an initialization of the CNN gives better results than doing it through the CNN own initialization., Los incendios forestales pueden tener un gran impacto en nuestro entorno y la naturaleza. Es por eso por lo que es interesante conocer su comportamiento. Nuestro proyecto quiere generar una reconstrucción 3D de los penachos del humo del fuego a partir de un número reducido de imágenes. Lo hacemos utilizando parametrización de cámaras para conocer la posición, forma y escala del objeto generando un Shape from Shilouette (SfS). La entrada serán dos fotos del humo tomadas a 90 grados la una de la otra mirando al mismo punto. La salida del SfS servirá para inicializar una red neuronal que convertirá el objeto en una densidad y suavizará los bordes. Se ha podido observar que usar un SfS como inicialización de la red neuronal da mejores los resultados que haciéndolo con la propia inicialización de la red neuronal., Els incendis forestals poden tenir un gran impacte en el nostre entorn i a la natura. És per això que és interessant conèixer el seu comportament. El nostre projecte pretén fer una reconstrucció en 3D dels plomalls del fum del foc a partir d’un nombre reduït d’imatges. Això ho farem utilitzant parametrització de càmeres per conèixer la posició, forma i escala de l’objecte generant un Shape from Silhouette (SfS). L’entrada seran dues fotografies del fum fetes a 90 graus una de l’altra ambdues mirant al mateix punt. La sortida del SfS servirà per inicialitzar una Xarxa Neuronal que convertirà l’objecte en una densitat i en suavitzarà les vores perquè s’assembli més a fum. S’ha observat que utilitzar un SfS com a inicialització de la Xarxa Neuronal dona millors resultats que fer-ho amb la pròpia inicialització de la Xarxa Neuronal.

Reverse engineering is a method of obtaining information about the geometry of an existing object. To obtain such information, among others: 3D laser scanning is used. The result of measurements using this method is a point cloud. The research examined the possibilities of using scanning data to analyze the technical condition of a historical building. Based on the inventoried point cloud, the deflections of beams and ceilings were determined. The course and width of scratches, cracks and defects were determined. It was found that the basic factor increasing the usefulness of the point cloud for various analyzes is its density, which depends on the accuracy of the scan performed. Thanks to a detailed point cloud, a digital three-dimensional model (digital twin) of the existing object was created and analyzed using computer methods. The aim of the presented research was to evaluate the use of reverse engineering to analyze the condition of a historic water tower. Moreover, the real possibilities and advantages of a relatively new measurement method were checked. The thesis was put forward that laser scanning and reverse engineering are effective methods supporting the assessment of the building condition. Based on the available literature and in situ tests, the problem was described and analyzes were carried out to assess the condition of the historic water tower building. Measurements with a 3D laser scanner and a tachymeter were also carried out in the field. Then, the digital model was used to measure the deflection and possible buckling of beams and ceilings on individual floors. The correctness of the formulated method for solving the problem was verified on the basis of tests of sample structural elements. It has been shown that it is possible to directly identify the elevations of the bottom of the ceiling elements and verify whether the existing deflections do not exceed the standard limits. As research has shown, a point cloud resulting from reverse engineer, This research was partially supported by the Departament de Recerca i Universitats de la Generalitat de Catalunya (Spain) (2021 SGR 01421 (GRBIO)). The authors also appreciate the fruitful discussions with members of the IEMAE and LABEDI for their contributions to this work. The research was carried out within the scope of works no. WZ/WB-IIL/4/2023 in Bialystok University of Technology and financed from the resources for science of Ministry of Science and Higher Education of Poland., Peer Reviewed, Objectius de Desenvolupament Sostenible::11 - Ciutats i Comunitats Sostenibles, Postprint (author's final draft)

Recent advances in robotics and nanotechnology resulted in a set of miniaturized Unmanned Aerial Vehicles (UAVs). Such small UAVs are envisioned to operate in hard-to-reach areas for enabling applications such as structural monitoring or content capturing. Towards showcasing this vision, we demonstrate a small UAV-supported setup for real-time autonomous generation of 3-Dimensional (3D) representations of static objects. In the setup, a small UAV (i.e., CrazyFlie 2.1) is envisioned to visit a set of locations, acting as a carrier and power source of a camera sensor. At each location, the sensor is expected to take a picture of the object and report it to the station. The station implements a pipeline for 3D reconstruction based on the pictures taken by the UAV., This work was funded by the MCIN/AEI/10.13039/01100011033/ FEDER/EU HoloMit 2.0 (PID2021-126551OB-C21) and MCIN/AEI/ 10.13039/501100011033 TRAINER-A (PID2020-118011GB-C21)., Peer Reviewed, Postprint (author's final draft)

Reverse engineering is a method of obtaining information about the geometry of an existing object. To obtain such information is used e.g. 3D laser scanning. The result of the measurements is a cloud of points (Fig. 1a). The paper presents the possibilities of using the cloud of points to analyze the technical condition of a historic building. Thanks to the cloud, it is possible to determine the deformations of structural elements, e.g. deformations of beams or ceilings. In tests over a long period of time, information about changes in deformation values can be obtained, and knowing the changes in loads (e.g. as a result of applying a test load), it is possible to measure the values of structure deflections resulting from these changes. The basic factor that increases the usefulness of a cloud of points for various measurements is its density, which depends on the accuracy of the scan. Thanks to the detailed cloud of points, it can be created a digital 3D model that will reflect the shape of an existing object. The results obtained during the measurement of the width and the depth of cracks are burdened with an error that is difficult to identify. However, the recorded accurate information about the location, number and course of cracks can be the basis for concluding the causes of their formation and the processes occurring in the structure. Over a long period of time, the recording of changes in the crack state allows conclusions about the stabilization or progress of the occurring damage. The main advantages of the method are speed, comprehensiveness, automatic measurement and high detail of the mapping of the structure geometry along with its damages., Peer Reviewed, Objectius de Desenvolupament Sostenible::11 - Ciutats i Comunitats Sostenibles, Postprint (published version)

Este documento presenta una investigación académica sobre la interoperabilidad en el modelado de información de construcción (BIM) para envolventes paramétricas, contribuyendo al ODS 11 Objetivo del Desarrollo Sostenible de Ciudades y Comunidades Sostenibles. El objetivo del estudio es desarrollar flujos de trabajo basados en un modelo geométrico de envolvente y un diccionario de elementos constructivos que garantice la interoperabilidad exitosa entre diferentes sistemas de software BIM, incluyendo Rhinoceros y el formato de uso abierto Industry Foundation Classes (IFC). Para lograr este objetivo, se ha llevado a cabo un análisis meticuloso de casos de estudio previamente expuestos en detalle. Se ha desarrollado un modelo geométrico de envolvente que representa de manera generalizada los componentes utilizados en la elaboración de fachadas paramétricas, y se ha creado un diccionario de elementos constructivos que incluye la identificación y definición de los diferentes elementos constructivos, como paneles, juntas, soportes, juntas de estanqueidad, herrajes, motores, sensores, iluminación y sistemas de control asociados. Cada uno de estos elementos ha sido definido de manera consistente y coherente en todos los sistemas relevantes. Además, se ha realizado la exportación a formato IFC y verificación de interoperabilidad de los modelos creados, comprobando que la geometría 3D se haya exportado correctamente y que se puedan leer los datos personalizados de cada elemento. Se ha utilizado el plugin de Geometry Gym para exportar directamente desde Rhinoceros, el flujo de trabajo con Rhino Inside para exportar desde Revit y el flujo de trabajo con Archicad Connection para exportar desde Archicad. En conclusión, los distintos flujos de trabajo y el diccionario de elementos constructivos desarrollados en este estudio contribuyen al ODS 11 de Ciudades y Comunidades Sostenibles al garantizar la interoperabilidad exitosa entre diferentes sistemas de software BIM para envolvente, Objectius de Desenvolupament Sostenible::11 - Ciutats i Comunitats Sostenibles

El objetivo de este trabajo es el de desarrollar un visualizador de modelos 3D de patrimonio cultural. Este visualizador es user-friendly, permitiendo a gente que se dedique al estudio del patrimonio y no sea experta en el uso de visualizadores por ordenador utilizar la herramienta con una curva de aprendizaje muy baja. Se han establecido una serie de funcionalidades, tanto de navegación como de interfaz gráfica que serían de gran utilidad para este tipo de usuarios. Entre estas funcionalidades se encuentran mostrar imágenes e información a medida que se navega sobre el modelo. El resultado final del trabajo es un visualizador implementado sobre la iglesia de Sant Quirze del Pedret plenamente funcional y con las funcionalidades mencionadas previamente. En primer lugar, se comentará el contexto, el alcance y los objetivos del trabajo, junto con las diferentes metodologías utilizadas. Se explicaría el sentido que tiene realizar el TFG y se le dará un contexto comparándolo con alternativas actuales. Posteriormente, se hablará sobre la implementación de la herramienta, tanto del sistema de navegación, como de la interfaz o de la carga de datos, junto con el entorno de desarrollo de la misma. Por ultimo se realizará unas conclusiones sobre el TFG y los objetivos conseguidos., The objective of this project is to develop a 3D model viewer for cultural heritage. This visualizer is user-friendly, allowing people who are dedicated to the study of heritage and who are not experts in the use of computer visualizers to use the tool with a very low learning curve. A set of functionalities have been established, both in terms of navigation and graphic interface, which will be very useful for this type of users. These functionalities include displaying images and information as the model is navigated. The final result of the work is a visualizer implemented on the church of Sant Quirze del Pedret fully functional and with the previously mentioned functionalities. First, the context, scope and objectives of the work will be discussed, together with the different methodologies used. The sense of doing the TFG will be explained and context will be given by comparing it with current alternatives. Subsequently, the implementation of the tool will be discussed, both the navigation system, the interface and the data loading, along with the development environment. Finally, some conclusions about the TFG and the objectives achieved will be made.

The discovery of conserved three-dimensional (3D) amino-acid patterns among a set of protein structures can be useful, for instance, to predict the functions of unknown proteins or for the rational design of multi-target drugs. There are several applications that perform a three-dimensional search of patterns in the structures of proteins. However, discovering conserved 3D patterns in a set of proteins with no other baseline patterns is a challenge. In this paper, we analyze and improve a state-of-the-art algorithm, 3D-PP, that implements this discovery. In this algorithm, the 3D patterns are detected and clustered using the root mean square deviation value, measured among each pair of 3D patterns (topological variability indicator). Even when 3D-PP deals with this task, the simultaneous processing of high amounts of proteins becomes a computational challenge with the size and the number of proteins to be evaluated. In this work, we present and analyze different shared memory parallel strategies of 3D-PP, using OpenMP. Those strategies improve the overall performance of the original implementation by reducing parallel load unbalance among threads and overall increasing parallelism. The results show significant performance improvements compared to the original version, achieving up to 13x speedup for a small number of proteins and 17.7× for a larger set., This research was funded by DICYT-USACH grant 5392102RP-AC and the Fondo Nacional de Desarrollo Cientifico y Tecnologico (FONDECYT) grants 1191133, Proyecto Fondecyt (Chile) 1220656 (MR-P), and by the Spanish Government (Grants PCI2021-121964 - TEXTAROSSA, PID2019-107255GB-C21/AEI/10.13039/501100011033, PID2019-107255GB-C22 - UPC-COMPUTACION DE ALTAS PRESTACIONES VIII and CEX2021-001148-S), and by Generalitat de Catalunya (2021 SGR 01007). Additional tests were performed in the cluster obtained with the grant CONICYT-FONDEQUIP-EQM160063., Peer Reviewed, Postprint (author's final draft)

We introduce Structured 3D Features, a model based on a novel implicit 3D representation that pools pixel-aligned image features onto dense 3D points sampled from a parametric, statistical human mesh surface. The 3D points have associated semantics and can move freely in 3D space. This allows for optimal coverage of the person of interest, beyond just the body shape, which in turn, additionally helps modeling accessories, hair, and loose clothing. Owing to this, we present a complete 3D transformer-based attention framework which, given a single image of a person in an unconstrained pose, generates an animatable 3D reconstruction with albedo and illumination decomposition, as a result of a single end-to-end model, trained semi-supervised, and with no additional postprocessing. We show that our S3F model surpasses the previous state-of-the-art on various tasks, including monocular 3D reconstruction, as well as albedo & shading estimation. Moreover, we show that the proposed methodology allows novel view synthesis, relighting, and re-posing the reconstruction, and can naturally be extended to handle multiple input images (e.g. different views of a person, or the same view, in different poses, in video). Finally, we demonstrate the editing capabilities of our model for 3D virtual try-on applications., Peer Reviewed, Postprint (published version)

Este proyecto tiene como finalidad realizar un estudio de la implantación de un láser 3D de visión artificial. Tiene la finalidad de realizar un control de calidad en cualquier proceso industrial y así poder detectar si los objetos están en buen estado, es decir, carecen de defectos. Para la realización de este estudio se ha utilizado un láser 3D de la empresa Cognex. Para la iluminación, se ha utilizado la propia iluminación aportada por el láser. Para el procesado de imágenes se ha utilizado el software In-Sight Vision Suite proporcionado por Cognex que es un software compatible al láser., Aquest projecte té com a finalitat realitzar un estudi de la implantació d’un làser 3D de visió artificial. Té la finalitat de realitzar un control de qualitat en qualsevol procés industrial y així poder detectar si els objectes estan en bon estat, es a dir, no tenen defectes. Per a la realització d’aquest estudi s’ha utilitzat un làser 3D de l’empresa Cognex. Per a la il•luminació, s’ha utilitzat la il•luminació pròpia aportada pel làser. Per al processat d’imatges s’ha utilitzat el software In-Sight Vision Suite proporcionat per Cognex que es un software compatible al làser., The purpose of this project is to carry out a study of the implantation of a 3D laser for artificial vision. Its purpose is to carry out quality control in any industrial process and thus be able to detect if the objects are in good condition, that is, they have no defects. To carry out this study, a 3D laser from the company Cognex has been used. For illumination, the illumination provided by the laser itself has been used. For image processing, it has been used the software In-Sight Vision Suite provided by Cognex which is compatible with the laser.

Aquest document presenta un informe exhaustiu sobre un projecte destinat a desenvolupar un procés automatitzat per a la creació d'animacions 2D a partir de models 3D utilitzant Blender. L'objectiu principal del projecte és millorar les tècniques existents i reduir la necessitat que els artistes realitzin tasques repetitives en el procés de producció d'animació. El projecte implica el disseny i desenvolupament d'un complement per a Blender, programat en Python, que es va desenvolupar per ser eficient i reduir les tasques intensives en temps que solen caracteritzar algunes etapes en el procés d'animació. El complement suporta tres estils específics d'animació: l'art de píxel, "cel shader", i "cel shader" amb contorns, i es pot expandir per suportar una àmplia gamma d'estils. El complement també és de codi obert, permetent una major col·laboració i potencials contribucions per part de la comunitat. Malgrat els problemes trobats, el projecte ha estat exitós en aconseguir els seus objectius, i els resultats mostren que el complement pot aconseguir resultats similars als adquirits amb eines similars i animació tradicional. El treball futur inclou mantenir el complement actualitzat amb les últimes versions de Blender, publicar-lo a GitHub i mercats de complements de Blender, així com afegir nous estils d'art., This document presents a comprehensive report on a project aimed at developing an automated process for creating 2D animations from 3D models using Blender. The project's main goal is to improve upon existing techniques and reduce the need for artists to do clerical tasks in the animation production process. The project involves the design and development of a plugin for Blender, coded in Python, which was developed to be efficient and reduce time-intensive tasks that usually characterise some stages in the animation process. The plugin supports three specific styles of animation: pixel art, cel shading, and cel shading with outlines, and can be expanded to support a wider range of styles. The plugin is also open-source, allowing for greater collaboration and potential contributions from the community. Despite the challenges faced, the project was successful in achieving its goals, and the results show that the plugin could achieve results similar to those acquired with similar tools and traditional animation. The future work includes keeping the plugin up-to-date with the latest versions of Blender, publishing it on GitHub and Blender plugin markets, as well as adding new art styles.

This work deals with the automatic 3D reconstruction of objects from frontal RGB images. This aims at a better understanding of the reconstruction of 3D objects from RGB images and their use in immersive virtual environments. We propose a complete workflow that can be easily adapted to almost any other family of rigid objects. To explain and validate our method, we focus on guitars. First, we detect and segment the guitars present in the image using semantic segmentation methods based on convolutional neural networks. In a second step, we perform the final 3D reconstruction of the guitar by warping the rendered depth maps of a fitted 3D template in 2D image space to match the input silhouette. We validated our method by obtaining guitar reconstructions from real input images and renders of all guitar models available in the ShapeNet database. Numerical results for different object families were obtained by computing standard mesh evaluation metrics such as Intersection over Union, Chamfer Distance, and the F-score. The results of this study show that our method can automatically generate high-quality 3D object reconstructions from frontal images using various segmentation and 3D reconstruction techniques., Postprint (published version)

6G and beyond networks will merge communication and computation capabilities in order to adapt to changes. As they will consist of many sensors gathering information from its environment, new schemes for managing these large amounts of data are needed. For this purpose, we review Over the Air (OTA) computing in the context of estimation and detection. For distributed scenarios, such as a Wireless Sensor Network, it has been proven that a separation theorem does not necessarily hold, whereas analog schemes may outperform digital designs. We outline existing gaps in the literature, evincing that current state of the art requires a theoretical framework based on analog and hybrid digital-analog schemes that will boost the evolution of OTA computing. Furthermore, we motivate the development of 3D networks based on OTA schemes, where satellites function as sensors. We discuss its integration within the satellite segment, delineate current challenges and present a variety of use cases that benefit from OTA computing in 3D networks., This work has received funding by the Spanish ministry of science and innovation under project IRENE (PID2020-115323RB-C31) funded by MCIN/AEI/10.13039/501100011033., Peer Reviewed, Postprint (author's final draft)

© 2009 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes,creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., We present a complete automatic system to obtain a realistic 3D avatar reconstruction of a person using only a frontal RGB image. Our proposed workflow first determines the pose, shape and semantic information from the input image. All this information is processed to create the skeleton and the 3D skinned textured mesh that forms the final avatar. We use a specific head reconstruction method to correctly match our final mesh to a realistic avatar. Our pipeline focuses on three main aspects: automation of the process, identification of the person, and usability of the avatar., Postprint (author's final draft)

One of the most essential interactions in Virtual Reality (VR) is the user’s ability to move around and explore the virtual environment. The design of the navigation technique plays a crucial role in the user experience since it determines key usability aspects. VR devices allow for an immersive exploration of 3D worlds, but navigation in VR is challenging for many users, due to potential usability issues related to specific VR controllers, user skills, and motion sickness. Although hundreds of interaction techniques have been proposed for this task, VR navigation still poses a high entry barrier for many users. In this paper we argue that adapting the navigation technique to its context of use can lead to substantial improvements in navigation usability and accessibility. The context of use includes the type of scene, the available physical space, as well as the profile of the user. We present a test platform to facilitate the design and fine-tuning of interaction techniques for 3D navigation. We focus on mainstream VR devices (headsets and controllers) and support the most common navigation metaphors (walking, flying, teleportation). The key idea is to let developers specify, at runtime, the exact mapping between user actions and locomotion changes, for any of the supported metaphors. Such mappings are described by a collection of parameters (e.g. maximum speed) whose values can be adjusted interactively through a GUI, or be provided by user-defined code which can be edited at runtime. Feedback obtained from developers suggests that this approach can be used to quickly adapt the navigation techniques to various people including persons with no previous 3D navigation skills, elderly people, and people with disabilities, as well as to the type, size and semantics of the virtual environment., This work has been funded by MCIN/AEI/10.13039/501100011033/FEDER ‘‘A way to make Europe’’. Pedret model partially funded by EU Horizon 2020, JPICH Conservation, Protection and Use initiative (JPICH-0127) and the Spanish Agencia Estatal de Investigación, grant PCI2020-111979 Enhancement of Heritage Experiences: the Middle Ages; Digital Layered Models of Architecture and Mural Paintings over Time (EHEM), Peer Reviewed, Postprint (published version)

La revolución experimentada por el uso creciente de la geoinformación 3D en la ingeniería geológica, gracias a los avances tecnológicos, que progresivamente han facilitado una disponibilidad de técnicas de adquisición de datos y de herramientas para su análisis. Junto a los avances experimentados en la disponibilidad de recursos de computación y almacenamiento en la nube, han impulsado la expansión de numerosas aplicaciones que permiten la difusión de la geoinformación 3D en la web. A pesar de estos avances científico-técnicos, es común ver los datos 3D simplificados en imágenes estáticas 2D, perdiendo parte su potencialidad comunicativa. El objetivo de esta comunicación es presentar el diseño y las funcionalidades de unos visores web 3D de código abierto como una herramienta de sencilla utilización con la que interactuar y visualizar los resultados. Así mismo, estamos ensayando las posibilidades de la realidad mixta como una experiencia plenamente inmersiva para lograr una mejor comunicación de la geoinformación 3D., Peer Reviewed, Postprint (published version)

Estimating accurate 3D pose and shape from a 2D image is an inherently difficult problem. Part of the difficulty arises from the ambiguity of potential solutions based solely on geometric features. The fields of computer vision and artificial intelligence are particularly suited to finding a solution to this problem although they have primarily focused on pose recovery, leaving shape as an afterthought. This thesis explores adaptations of and extensions to a recent human mesh recovery framework that showed a significant improvement on shape metrics compared to a very popular real-time pose and shape estimator. The framework employs a multi-stage process, refining pose first, and then shape, through a non-differentiable mesh deformation process. A differentiable alternative to these deformation steps was proposed. In support of this effort, a dataset was compiled which indexes some of the most popular 2D and 3D human datasets and provides a common access format. The mesh recovery framework was retrained using this dataset, which incorporated an order of magnitude more samples than the dataset used to train the published framework. The new weights achieved the same levels of performance as the published weights, despite having less reliable ground-truth annotations. In addition, a multi-layer perceptron that has demonstrated state-of-the-art performance at pose parameter regression was trained, using millions of ground-truth 3D human meshes, to correct perturbations in shape and pose. Training techniques and methods of interfacing this network to the mesh recovery framework have been investigated and documented.

En aquest projecte ens hem proposat desenvolupar un sistema de visualització de dades en forma de gràfica en un entorn 3D per facilitar la comprensió d'aquestes dades per a qualsevol usuari. Per fer-ho, hem creat una aplicació de creació de gràfiques en 3D mitjançant tecnologies de realitat virtual immersiva, fent ús del motor de creació de videojocs Unity3D i de diferents perifèrics necessaris per executar l'aplicació, com per exemple el casc de realitat virtual HTC Vive Pro Eye. L'aplicació permet crear fins a un total de tres tipus de gràfiques diferents (de punts, de línies i de barres) i personalitzar-ne diferents aspectes. Les dades entren al sistema mitjançant un fitxer .CSV i l'aplicació està preparada per acceptar dades tant numèriques com categòriques. Per validar l'aplicació i ampliar la recerca prèvia hem dut a terme tests d'usuaris, que ens han permès valorar la percepció dels diferents valors de les gràfiques de barres en un entorn de realitat virtual immersiva. Els resultats ens han permès valorar diferent informació sobre la percepció dels usuaris, els errors que han comes, l'evolució dels errors segons el temps, la gravetat dels errors que cometien, la distribució de les mitjanes dels errors, la desviació estàndard dels errors, la mediana dels errors i tota la informació possible sobre la mirada dels usuaris durant tot l'experiment. En conclusió, hem pogut comprovar que aquesta aplicació ens permet mostrar de manera correcta les dades en una gràfica de dades en realitat virtual 3D immersiva., In this project, we have developed a data visualisation system in terms of graphics in a 3D environment to facilitate the comprehension of the data for any user. To do so, we implemented a 3D graphics creation application using immersive virtual reality applications, making use of the video game tools like Unity3D and different accessories that are needed to run our application, as for instance, the virtual reality helmet HTC Vive Pro Eye. The application allows the creation of three different graphic types (dot, lines and bar) and to custom several aspects of them. The input data is provided as a .CSV file and the application is prepared to accept either numeric or categorical data. To validate our application and extend previous research we have done user test that have allowed to validate the perception of different values on the bar plots in the immersive virtual reality environment. The results allowed us to consider information about the perception of the users, the errors, the evolution of the errors according to the time, the strength of these errors, its mean, standard deviation and median distribution and all the information of the eye tracking of the user during the whole experiment. In conclusion, we have been able to check that this application shows properly the data in a 3D immersive virtual reality environment

Terrestrial Laser Scanners, also known as LiDAR, are often equipped with color cameras so that both infrared and RGB values are measured for each point sample. High-end scanners also provide panoramic High Dynamic Range (HDR) images. Rendering such HDR colors on conventional displays requires a tone-mapping operator, and getting a suitable exposure everywhere on the image can be challenging for 360° indoor scenes with a variety of rooms and illumination sources. In this paper we present a simple-to-implement tone mapping algorithm for HDR panoramas captured by LiDAR equipment. The key idea is to choose, on a per-pixel basis, an exposure correction factor based on the local intensity (infrared reflectivity). Since LiDAR intensity values for indoor scenes are nearly independent from the external illumination, we show that intensity-guided exposure correction often outperforms state-of-the-art tone-mapping operators on this kind of scenes., This work has been partially funded by the Spanish Ministry of Economy and Competitiveness and FEDER under grant TIN2017-88515-C2-1-R, the Romanesque Pyrenees, Space of Artistic Confluences II (PRECA II) project (HAR2017-84451-P, UB) and by EU Horizon 2020, JPICH Conservation, Protection and Use initiative (JPICH-0127) and the Spanish Agencia Estatal de Invesigación (grant PCI2020-111979). The authors would like to thank Prof. Xavier Pueyo for his helpful comments., Peer Reviewed, Postprint (published version)

Laser scanners enable the digitization of 3D surfaces by generating a point cloud where each point sample includes an intensity (infrared reflectivity) value. Some LiDAR scanners also incorporate cameras to capture the color of the surfaces visible from the scanner location. Getting usable colors everywhere across 360° scans is a challenging task, especially for indoor scenes. LiDAR scanners lack flashes, and placing proper light sources for a 360° indoor scene is either unfeasible or undesirable. As a result, color data from LiDAR scans often do not have an adequate quality, either because of poor exposition (too bright or too dark areas) or because of severe illumination changes between scans (e.g. direct Sunlight vs cloudy lighting). In this paper, we present a new method to recover plausible color data from the infrared data available in LiDAR scans. The main idea is to train an adapted image-to-image translation network using color and intensity values on well-exposed areas of scans. At inference time, the network is able to recover plausible color using exclusively the intensity values. The immediate application of our approach is the selective colorization of LiDAR data in those scans or regions with missing or poor color data., This work has been partially funded by the Spanish Ministry of Economy and Competitiveness and FEDER under grant TIN2017-88515-C2-1-R, the Romanesque Pyrenees, Space of Artistic Confluences II (PRECA II) project (HAR2017- 84451-P, UB) and by the EU Horizon 2020, JPICH Conservation, Protection and Use initiative (JPICH-0127) and the Spanish Agencia Estatal de Invesigación (grant PCI2020-111979)., Peer Reviewed, Postprint (published version)

Every single day, researchers and developers find new ways where VR devices could be used to provide considerable benefits to a variety of fields. The cheap entry price, alongside with the lack of real-world consequences, makes it a really attractive prospect in fields such as medicine and vehicle driving simulations. Moreover, the current COVID-19 crisis is affecting training scenarios that would otherwise have been done in person, due to both the lack of resources and the risk of contagion, especially in the medical field. This project aims to provide a portable system that allows nursing personnel to create, configure, and play training sessions in an immersive virtual environment. Thanks to the accessible Oculus Quest, it provides an environment for easily creating training sessions, and experiencing these sessions in an immersive experience through the use of a VR headset and controllers., Cada dia, investigadors i desenvolupadors troben noves maneres en les quals els dispositius VR podrien ser utilitzats per a beneficiar a una varietat de camps. El baix cost inicial, conjuntament a la mancança de risc real, fa que aquest prospecte resulti atractiu per a certs camps com el de la medicina i el de les simulacions de conducció de vehicles. A més a més, la situació actual del COVIC-19 esta afectant a l'entrenament de personal que s'hauria d'haver realitzat en persona, ja sigui per culpa d'una mancança de recursos o el risc de contagi, especialment en el camp de la medicina. L'objectiu d'aquest projecte és proporcionar un sistema portàtil que permeti al personal d'infermeria crear, configurar, i reproduir sessions d'entrenament en un ambient virtual. Gracies a l'accessibilitat de l'Oculus Quest, es proporciona un entorn per a la fàcil creació de sessions d'entrenament, i per a experimentar aquestes sessions en una experiència immersiva emprant els cascs i controladors VR.

Photogrammetry is one of the widest techniques used to monitor terrain changes which occur due to natural process and geological natural risk zones. In order to carry out terrain monitoring, it is necessary to eliminate all the non-ground elements. One of the most variable elements in this monitoring is the presence of vegetation, which obscures the ground and can significantly mislead any multitemporal analysis to detect terrain changes. Therefore, the focus of this paper is about how best to filter the vegetation to attain an accurate reading of the terrain. There are several methods to filter it based on colourising an excessive greenness vegetation index or non-visible channels as the IR in the well-known index NVDI. However, achieving this kind of information is not always possible because its high cost. Instead this channel we can add new information using the HSV colour space obtained from the RGB information. In this paper, we propose a double possibility, on one hand work wit h RGB+HSV for a supervised segmentation on images. On the other, to use excessive greenness vegetation indices and RGB+HSV for the segmentation of point clouds. The results shown that the use of additional channels HSV can significantly improve the segmentation in both studies, and therefore render a much more accurate assessment of the underlying terrain, Peer Reviewed, Postprint (published version)

In this work, we perform a study of different network architectures for 3D garment reconstruction. In the study, five networks are compared both quantitatively and qualitatively. The dataset used to train each network is CLOTH3D, which consists of six different garment types: tshirt, top, jumpsuit, dress, skirt, and trousers. The baseline architecture is introduced in the CLOTH3D paper. Each of the other four networks are defined by introducing modifications to this baseline: 1) applying central difference convolution (CDC) to GNNs, 2) new pooling based on spectral clustering, 3) applying octave convolution to GNNs, and 4) combining the CDC and pooling networks (CDC-pool). We show that the CDC and pooling networks independently outperform the baseline, octave, and CDC-pool networks quantitatively. As well as, the pooling network is best suited for modelling the complex dynamics present in the dress and skirt garment types, while there is little difference between the networks for the top, tshirt, trousers, and jumpsuit garment types.

© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting /republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works, In this paper we introduce SMPLicit, a novel generative model to jointly represent body pose, shape and clothing geometry. In contrast to existing learning-based approaches that require training specific models for each type of garment, SMPLicit can represent in a unified manner different garment topologies (eg from sleeveless tops to hoodies and to open jackets), while controlling other properties like the garment size or tightness/looseness. We show our model to be applicable to a large variety of garments including T-shirts, hoodies, jackets, shorts, pants, skirts, shoes and even hair. The representation flexibility of SMPLicit builds upon an implicit model conditioned with the SMPL human body parameters and a learnable latent space which is semantically interpretable and aligned with the clothing attributes. The proposed model is fully differentiable, allowing for its use into larger end-to-end trainable systems. In the experimental section, we demonstrate SMPLicit can be readily used for fitting 3D scans and for 3D reconstruction in images of dressed people. In both cases we are able to go beyond state of the art, by retrieving complex garment geometries, handling situations with multiple clothing layers and providing a tool for easy outfit editing. To stimulate further research in this direction, we will make our code and model publicly available at http://www.iri.upc.edu/people/ecorona/smplicit/., his work is supported in part by an Amazon Research Award and by the Spanish government with the projects HuMoUR TIN2017-90086-R and María de Maeztu Seal of Excellence MDM2016-0656. Gerard Pons-Moll is funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) - 409792180 (Emmy Noether Programme, project: Real Virtual Humans), Peer Reviewed, Postprint (published version)

Aquest projecte consisteix en la creació d’un Lleó en 3D, que s’entregarà en format de vídeo, on es mostrarà l’animal en una animació en la que es puguin apreciar tots el processos que s’efectuaran durant el desenvolupament del projecte. Es realitzaran les tasques des del plantejament de la idea inicial fins al estudi de l’anatomia de l’animal, passant per tota la producció, des del modelatge, fins a la animació del moviment tant de l’animal com dels seus pèls, generant tots els efectes visuals necessaris fins aconseguir el vídeo final, per poder crear un producte de referència quant a tècniques empleades i al tipus de resultat que es pot aconseguir amb pocs recursos i poc de temps per invertir. En la part principal del projecte que es tracta de la postproducció, s’utilitzaran diverses tècniques avançades habituals en l’àmbit professional i es desenvoluparan sistemes específics per crear les diferents situacions necessàries en 3D. Les tècniques que s’utilitzaran durant el procés consisteixen en el modelatge 3D, texturització, animació i grooming de l’animal. Per tal d’aconseguir que el resultat sigui el més realista possible, es partirà de la cerca d’informació veraç, que s’estudiarà, analitzarà i s’aplicarà durant el procés de creació del model 3D

Este trabajo se centra en el desarrollo de una experiencia interactiva virtual, pensada desde el principio para poder ser disfrutada usando cascos de realidad virtual de tipo Oculus Rift o HTC Vive. Se desarrolla desde la voluntad de adquirir más conocimientos de esta tecnología y de los procesos que hay que seguir para conseguir un producto final que sea apto para personas que no tengan muchos conocimientos técnicos en la materia. Además está centrado en adquirir conocimientos más profundos de optimización del proceso de creación de entornos virtuales en cuanto a modelado, texturizado e iluminación, ya que son fundamentales para crear una sensación mayor de realidad

El projecte consisteix en la creació d'una videodansa en què les articulacions de la ballarina es tracen amb diferents grafismes amb animació digital, a més a més, d’aplicar tècniques com la integració d'elements 3D. La peça s'emmarca dins del món de la dansa on els moviments de la ballarina queden plasmats amb formes geomètriques. Tanmateix, aquests desplaçaments del cos amb la combinació de l'animació digital es crea una il·lusió de moviment i una immersió amb la geometria euclidiana. La videodansa s'ha treballat des del plantejament amb la idea inicial, la direcció de tota la preproducció i producció i sobretot s'ha posat un gran interès amb la postproducció per tal d'aconseguir el vídeo el més professional possible. Les tècniques que s'han utilitzat durant el procés per crear les formes geomètriques amb la integració d'imatges reals han consistit en la creació de motion graphics de forma conjunta amb la tècnica de rotoscòpia i també la integració 3D dels elements i la simulació d'aquests. Per realitzar el treball també s'han portat a terme moltes hores d'anàlisi i d'investigació. En conjunt el procés ha estat complicat amb tots els entrebancs que han anat sorgint durant el projecte, sí que és cert que tots han tingut la corresponent solució. El resultat del projecte és molt satisfactori, ja que s'han pogut aconseguir els objectius plantejats a l'inici, obtenint una experiència molt gratificant

Este proyecto trata el análisis de las técnicas cinematográficas utilizadas para coreografiar una batalla entre criaturas de tamaños colosales, el objetivo será la realización de un trabajo práctico de animación básica basándonos en el análisis hecho previamente. Este proyecto abarca desde la creación de un guión mediante el análisis de referentes de la industria, la creación de un storyboard y un box-sizing de dicho guión, dependiendo del tiempo del que disponga se procedería con el modelado y rigging de los personajes y el entorno. Para ello he analizado muchas películas en las que aparecen grandes criaturas ya sean orgánicas o mecánicas, para analizar el comportamiento y el tratamiento que se les da a cada una. También me he fijado en los tipos de planos, los movimientos de cámara y cómo estos afectan a la transmisión de unos valores y emociones u otros. Los referentes para este trabajo han sido desde las películas de godzilla de la década de los 90 hasta las últimas entregas de Pacific Rim, decir también que no solo me he fijado en los aspectos positivos de las películas, he visto películas malas, de bajo presupuesto y con múltiples errores para poder así extraer conclusiones sobre los errores para no cometerlos a la hora de hacer yo mi cortometraje. Respecto a la metodología, he utilizado Trello para poder organizarme y ver de manera sencilla la carga de trabajo y la Técnica Pomodoro1 para una mejor gestión del tiempo de trabajo. Los resultados de este proyecto han sido satisfactorios, habiendo conseguido un resultado satisfactorio mediante la puesta en práctica de los conocimientos resultantes de la investigación anterior

Measuring nascent macromolecular synthesis in vivo is key to understanding how cells and tissues progress through development and respond to external cues. Here we perform in vivo injection of alkyne- or azide-modified analogs of thymidine, uridine, methionine, and glucosamine to label nascent synthesis of DNA, RNA, protein, and glycosylation. Three-dimensional volumetric imaging of nascent macromolecule synthesis was performed in axolotl salamander tissue using whole-mount click chemistry-based fluorescent staining followed by light sheet fluorescent microscopy. We also developed an image processing pipeline for segmentation and classification of morphological regions of interest and individual cells, and we apply this pipeline to the regenerating humerus. We demonstrate our approach is sensitive to biological perturbations by measuring changes in DNA synthesis after limb denervation. This method provides a powerful means to quantitatively interrogate macromolecule synthesis in heterogenous tissues at the organ, cellular, and molecular levels of organization, Peer Reviewed, Postprint (published version)

Pedestrian detection algorithms have always revolved around RGB scene information, but relying solely on it can be dangerous in situations where conventional cameras don't capture reality properly. For this reason, in recent years, many researchers have studied other alternatives that complement these existing techniques, such as the use of ultrasonic sensors or radars, that provide more reliable information in those situations. Another approach is to use LIDAR sensors, which map reality into point clouds using pulses of light. However, there are few studies that propose pedestrian detection techniques using only the data provided by a LIDAR. In this thesis, we explore this approach through the design and implementation of a pedestrian detection system in 3D point clouds. To do so, we train the PointNet++ point cloud classification network in order to demonstrate that the 3D geometric information of a scene is essential for the neural network to learn properly. Specifically, to carry out supervised training we need to generate a pedestrian and non-pedestrian ground truth in point clouds, so we have designed a semi-automatic labeling system based on the detection in RGB images and the subsequent transfer of these labels to the 3D domain. As a result, we train PointNet++ and test its performance on an outdoor dataset, obtaining outstanding results of up to 99.4% of accuracy and 98.6% of recall. With these results we are firmly corroborating the hypothesis stated in the thesis that 3D geometric information is essential for a neural network to learn to detect pedestrians in outdoor scenes. Not only that, we also surpass the results provided by a conventional detector in RGB images: YOLO, which provides a 48% of recall in the same dataset, thus proving that geometric information should not be an alternative in these systems, but a must., Los algoritmos de detección de peatones basan sus principios en la información RGB de las escenas, lo que los convierte en sistemas poco fiables en condiciones atmosféricas desfavorables. Por este motivo, durante los últimos años se han estudiado alternativas que complementen este tipo de datos, desde el uso de sensores ultrasónicos y radares hasta los sensores LIDAR, que emiten pulsos de luz para mapear el entorno en forma de nubes de puntos tridimensionales. Algunos investigadores optan por combinar la información RGB proporcionada por las cámaras convencionales con la información geométrica que proporciona un LIDAR, pero hay muy pocos enfoques en que se explore únicamente la importancia de la información geométrica. Por este motivo, en esta tesis se presenta un sistema para detectar peatones en nubes de puntos 3D usando PointNet++ como clasificador en datasets de exteriores, con el objetivo de demostrar la relevancia que la información geométrica tiene en la percepción del entorno. Con tales fines, se diseña un sistema de etiquetaje en nubes de puntos semiautomático basado en la detección de peatones y no peatones en imágenes RGB y en la posterior transferencia de estas etiquetas al dominio 3D. Una vez generado este groundtruth de peatones y no peatones en nubes de puntos, lo usamos para entrenar a PointNet++, obteniendo unos resultados excelentes. La red neuronal proporciona una precisión y un recall de 99.4% y 98.6%, respectivamente, muy por encima del 48% de recall proporcionado por YOLO, un detector en imágenes RGB convencional, con los mismos datasets. De este modo, en esta tesis se prueba rotundamente que la información geométrica es un elemento esencial para comprender los elementos que nos rodean., L'exploració en el camp de la detecció de vianants en imatges RGB és un àmbit molt treballat desde fa anys, però dependre únicament de la informació fotomètrica d'una escena pot ser perillós en algunes situacions. Per aquest motiu, cada cop hi ha més interès en la detecció basada en altres tipus de dades: sensors ultrasònics, radars o els LIDAR, un tipus de sensor que fa un mapeig de la realitat en forma de núvols de punts mitjançant polsos de llum. La informació geomètrica que proporciona un núvol de punts pot ser molt útil en situacions on les imatges no són prou fiables (foscor, mal temps), i és per això que en aquesta tesi es presenta un sistema de detecció de vianants en núvols de punts 3D. Per demostrar la importància de la informació geomètrica a l'hora d'entendre el nostre entorn, es proposa entrenar PointNet++, una xarxa de classificació en núvols de punts, per tal de verificar la capacitat d'aquesta xarxa d'aprendre a diferenciar persones en datasets d'exteriors únicament a partir de la informació geomètrica. Amb aquest propòsit, es dissenya un sistema per generar un ground truth de vianants i no vianants a núvols de punts, mitjançant un etiquetatge semiautomàtic basat en la detecció en imatges RGB i la transferència d'aquestes deteccions als núvols de punts corresponents. D'aquesta manera, podem utilitzar un dataset lo suficientment extens i variat com perquè PointNet++ pugui aprendre d'ell. Un cop entrenada, la xarxa neuronal proporciona uns resultats excel·lents, arribant a una precisió i recall de fins a 99.4% i 98.6%, respectivament. Aquests valors superen totalment les deteccions de YOLO, un detector en imatges RGB convencional, provant així la hipòtesi definida a la tesi que la informació geomètrica és un element imprescindible per a reconèixer el nostre entorn.

Bone regeneration constitutes a major field within tissue engineering and it is receiving widespread attention. Bone fractures are often difficult to heal. Traditional repair methods usually yield unsatisfactory results. One of the most recent techniques are the so-called scaffolds. These scaffolds are small porous pieces made of Calcium Phosphate. Once inserted in the bone fracture, they foster bone formation by, among other things, releasing Ca2+ ions. Pore geometry is believed to play a crucial role by helping to retain the ions in the scaffold. Therefore, determining the optimal pore geometry is key for a standardized production of scaffolds through, for example, 3D printing. Numerical simulation offers the possibility to test different designs at a very low cost. A numerical model that would be able to reproduce the behaviour of the scaffolds would be a very powerful tool. In this thesis, we have put forward a simple model to reproduce the evolution of the concentration. Using the advection-diffusion equation we have modelled the evolution of Ca2+ ions in two different scaffolds: a foam scaffold (random structure and difficult to produce)and 3D printed scaffold (ordered structure and easy to produce). The performance of these scaffolds was tested in vivo by Barba et al. (2018). The simulations that we have performed seem to match the experimental results: higher concentrations appear in the foam scaffold. Additionally, we have tested the evolution of the concentration under three different velocity fields. For the foam scaffold, the best results are obtained with a simple uniaxial flow, normal to one of the faces of the scaffold. As for the 3D scaffold, the highest concentration is obtained with two orthogonal flows and normal to the faces. The implementation of the equation has been done using FEniCS, an open-source platform to solve PDEs.

El presente trabajo tiene como objeto presentar el proyecto de innovación educativa Buildings 360º. Proyecto que toma el relevo de iniciativas educativas anteriores, pretendiendo desarrollar una novedosa plataforma inmersiva que permitirá consultar las obras de construcción a lo largo del tiempo. Plataforma que abogará por el uso de imágenes 360º en contraposición de enfoques más extendidos basados en costosos modelos 3D, dando lugar a un sistema de bajo coste, muy intuitivo, capaz de gestionar grandes bloques de información y fácil de implementar por usuarios no expertos. Este sistema se usará como complemento docente en la asignatura de Construcción I y Construcción II, permitiendo al alumnado consultar cada una de las fases de una obra, sus materiales y sistemas constructivos., We present here the innovative education project Buildings 360⁰. This project follows previous educative initiatives. The project goal is to develop a novel immersive platform that will enable students to view real construction works along time. This platform will use 360 images in contrast to the widespread and expensive 3D models resulting into a low-cost system which is at the same time very intuitive, allows to manage large amounts of data and is easy to implement by non-expert users. This system will be used as a complementary material in the subjects of Construction I and II and will enable students to look up each of the phases of the construction work, the materials and the construction systems., Este trabajo ha sido financiado por la Universidad Politécnica de Madrid a través del proyecto de innovación educativa titulado “Integración de enfoques 360º en el aprendizaje de la construcción” (Buildings 360º) con identificador IE1920.0302 y otorgado en la convocatoria 2019/2020.De manera adicional los autores quieren mostrar su agradecimiento al resto del equipo docente que intervienen en dicha iniciativa, así como a Lisa Hillerbrand (becaria contratada con cargo al proyecto) y al Departamento de Construcción y Tecnología Arquitectónicas por el apoyo prestado (www.dcta.upm.es)., Peer Reviewed

Desde el año 2013, el índice de incidencia del sector de la construcción en España ha crecido un 33%. La construcción lidera el aumento de la siniestralidad respecto a otras industrias. Los sistemas educativos existentes en prevención de riesgos laborales no logran el objetivo deseado mitigar este problema, ni se tratan con la debida importancia. Por lo que nos vemos obligados a fomentar una cultura preventiva con un sistema educativo interactivo y efectivo. En la búsqueda de herramientas que permitan educar; proporcionando un entorno de entrenamiento completo e inmersivo, la literatura propone el uso de Serious Game. El propósito de este trabajo es elaborar una herramienta interactiva e inmersiva, que permita capacitar a los técnicos en prevención de riesgos laborales (TPRL). Con esta herramienta se busca mejorar sus habilidades en las visitas de inspección de obra. La herramienta desarrollada se basa en un flujograma de funciones de un TPRL. Este instrumento de formación ha sido desarrollado en el motor de video juegos Unity 3D. Mediante un modelo tridimensional de un proyecto de construcción y, con la implementación de objetos y acciones, recrea una visita real de inspección de obra. Utiliza el lenguaje de programación C#. Se realiza la validación de la herramienta desarrollada mediante un cuestionario que, aplicado a un panel de expertos, evalúa la experiencia del usuario., Since 2013, the incidence rate of the construction sector in Spain has grown by 33%. In respect to other industries, the construction industry is the leader in the accident rate increase. The existing educational systems for the prevention of occupational risks do not achieve the desired objective, nor are they treated with due importance. Therefore, we must promote a prevention culture with an interactive and effective educational system. In search of tools that allow educating by providing a complete and immersive training environment, the literature proposes the use of Serious Game. The purpose of this work is to elaborate an interactive and immersive tool, which allows training technicians in the prevention of labour risks (TPRL). With this tool, we seek to improve their skills in site inspection visits. We have based the developed tool on a flow chart with a TPRL's functions. We have developed this training tool in the Unity 3D video game engine. Through a three-dimensional model of a construction project and with the implementation of objects and actions, it recreates a real site inspection visit. It uses the C# programming language. Validation of the developed tool is done through a survey that, applied to a panel of experts, evaluates the user's experience.

Aquest treball de final de grau neix de l'estudi "Follower Behavior in a Virtual Enviroment", on l'usuari ha de realitzar diferents tasques en un entorn virtual que simula una estació de tren. Passats uns minuts sona l'alarma d'emergència de l'estació, la finalitat de l'estudi és analitzar el comportament dels participants en funció de com actuen els personatges autònoms de l'entorn virtual. L'objectiu del TFG és solucionar el problema provocat pel sistema de navegació utilitzat, Walk in Place. Amb el WiP, el participant es pot moure llargues distàncies dins de la simulació aixecant i baixant les cames, fent l'acció de caminar però sense moure's de la posició original. El problema és que la gent fa petits desplaçaments quan estan immersos i hi ha risc de col·lisió, ja que per l'equip de realitat virtual s'observa com es camina i inconscientment l'usuari fa aquests passos en la vida real. Durant el projecte, s'han investigat diferents mètodes per veure quina solució encaixaria millor. La solució final consisteix a estudiar el comportament dels personatges autònoms perquè realitzin una acció en concret, amb l'objectiu de reposicionar a l'usuari i assegurar-se que no col·lisiona amb l'entorn físic. Una part molt important d'aquest treball és mantenir el realisme que es vol aconseguir en entorns de realitat virtual, de tal manera que les accions dels personatges autònoms no siguin incoherents i es mantingui l'experiència immersiva del participant. Per desenvolupar aquest treball es necessitarà adquirir coneixements de la plataforma Unity i el llenguatge C#, a més a més d'estudis sobre realitat virtual., This undergraduate dissertation was born from the study "Follower Behavior in a Virtual Environment", where the user has to perform different tasks in a virtual environment that simulates a train station. After a while carrying out those tasks, the emergency alarm of the train station starts to rang. In any case the purpose of the study is to analyze both the behaviour of the participants and their interactions with the non player characters (NPCs) of the virtual environment. The goal of my undergraduate dissertation is to solve the issues caused by the navigation system used, Walk in Place. Such system entails that the participant can travel long distances in the simulation raising and lowering the legs, realizing the act of walking without moving from the original position. However the issue within WiP is that people make small displacements meanwhile immersed in the virtual reality therefore appearing a risk of collision. That effect is caused because using the virtual reality equipment the user sees how the avatar walks and unconsciously it reproduces those steps in real life. This project oversees different methods in order to select which solution would be the best. The final solution consists in programing the behaviour of the non player characters to perform specific actions, with the purpose of repositioning the user avoiding collisions with the physical environment, ensuring its safety. A key part of this dissertation is to conserve the realism pursued by virtual reality environments, in a way that the behaviour of the non player characters are not incoherent and conserve the immersive experience of the user. Finally in order for this work to be developed it will be required knowledge of the Unity platform, the language C# in addition to virtual reality studies., Este trabajo de final de grado nace del estudio "Follower Behavior in a Virtual Enviroment", donde el usuario tiene que realizar diferentes tareas en un entorno virtual que simula una estación de tren. Pasados unos minutos suena la alarma de emergencia de la estación, la finalidad del estudio es analizar el comportamiento de los participantes en función de cómo actúan los personajes autónomos del entorno virtual. El objetivo del TFG es solucionar el problema provocado por el sistema de navegación utilizado, Walk in Place. Con el WiP, el participante puede moverse largas distancias dentro de la simulación levantando y bajando las piernas, haciendo la acción de andar pero sin moverse de la posición original. El problema es que la gente hace pequeños desplazamientos cuando están inmersos y hay riesgo de colisión, debido a que por el equipo de realidad virtual se observa como se anda e inconscientemente el usuario hace estos pasos en la vida real. Durante el proyecto, se han investigado diferentes métodos para ver qué solución encajaría mejor. La solución final consiste en estudiar el comportamiento de los personajes autónomos para que realicen una acción en concreto, con el objetivo de reposicionar al usuario y asegurarse de que no colisiona con el entorno físico. Una parte muy importante de este trabajo es mantener el realismo que se quiere conseguir en entornos de realidad virtual, de tal manera que las acciones de los personajes autónomos no sean incoherentes y se mantenga la experiencia inmersiva del participante. Para desarrollar este trabajo se necesitará adquirir conocimientos de la plataforma Unity y el lenguaje C#, además de estudios sobre realidad virtual. 4

Este proyecto consiste en la creación de un vídeo de 1 minuto de duración que permita al autor mostrar cierto abanico de habilidades y destrezas relacionadas con la animación 3D y 2D, la edición de vídeo, el storytelling y la dirección de arte. A través de una marca personal ya creada, el vídeo servirá como porfolio del alumno en su carrera como artista de motion graphics, permitiéndole tener material para mostrar a la hora de ser contratado en un estudio o empresa del sector. Asimismo, servirá como herramienta para consolidar esta marca personal de cara a crear una audiencia en redes sociales que permita al alumno tener más visibilidad en internet y, por tanto, más posibilidades de llegar con su trabajo a más artistas del sector y a potenciales clientes. Con el fin de lograr estos objetivos, el vídeo constará de diferentes escenas de animación 3D y 2D de corta duración (1-5 segundos) aparentemente no relacionadas entre sí. Únicamente habrá un elemento común como hilo conductor del vídeo: el logotipo del alumno (creado anteriormente por él mismo) en la misma posición de la pantalla en todas las escenas, dando pie a diferentes representaciones e interpretaciones gráficas de este a lo largo del vídeo. La referencia principal para este proyecto ha sido la plataforma AvantForm, creada recientemente y presentada mediante un vídeo similar en el que el logotipo de la misma ha sido interpretado por los principales artistas que la componen, dando como resultado un variado abanico de estilos y conceptos conectados únicamente por la presencia de este elemento común a todas las escenas. Otros referentes son también artistas 3D como Ian Frederick o Philip Lück, que también han utilizado esta idea recientemente para promocionarse profesionalmente. La metodología que se seguirá en este proyecto es la llamada Kanban, una metodología ágil que permite al alumno tener un control de todas las fases del proyecto y que se enfoca en gestionar el tiempo de cara a las diferentes entregas

In the recent years, numerous Machine Learning investigations have been centered around the topic of point-target positioning and tracking, which rises the interest for algorithms such as k-NN to be implemented for said purposes using measurements from a radar network. In this Thesis, the k-NN algorithm has been introduced in a joint range-Doppler data association algorithm to provide a novel grid-based solution to distinguish between targets in multi-target scenarios using a network of fully-multistatic OFDM radar. A data fusion algorithm has also been developed to present positioning estimates using k-NN and velocity estimates with a closed-form expression. The results show good associating and fusion error, depending on the grid resolution defined and on the k parameter used. Moreover, the k-NN's great adaptability to all different target hypothesis makes the use of extended and complex prunning unnecessary in data association. Future work in this project would include taking a step towards dynamic data fusion with target-tracking algorithms, combined with the existing data association using k-NN, thus aiming to provide highly-reliable results in the real-life tracking of point-targets such as drones., Durante los últimos años, numerosas investigaciones de Machine Learning se han centrado alrededor del posicionamiento y el seguimiento de blancos, lo que aumenta el interés por algoritmos como el k-NN que pueden ser aplicados para estos propósitos mediante medidas extraídas de una red radar. En esta tesis, se ha introducido el k-NN en un algoritmo de asociación conjunto de rangos y Doppler para proporcionar una nueva solución para distinguir los blancos en escenarios multi-objetivo utilizando una red de radar OFDM completamente multiestática . También se ha desarrollado un algoritmo de fusión de datos para presentar estimaciones de posicionamiento con k-NN y estimaciones de velocidad con una expresión matemática de lazo cerrado. Los resultados muestran un error de asociación y fusión comprensibles, dependiendo de la resolución de la cuadricula definida y del parámetro k utilizado. Además, la gran adaptabilidad del k-NN a todas las diferentes hipótesis a considerar en la asociación de datos hace que la técnica no necesite filtrar datos utilizando algoritmos costosos en tiempo y en complejidad computacional. En el futuro, este trabajo avanzaría hacia la fusión de datos dinámica, utilizando algoritmos de seguimiento de blancos combinados con la asociación desarrollada con k-NN, con el objetivo de ofrecer resultados altamente fiables en el seguimiento de blancos como drones., En els darrers anys, nombroses investigacions de Machine Learning s'han centrat al voltant del posicionament i el seguiment de blancs, fet que augmenta l'interès en algoritmes com el k-NN que poden ser aplicats per aquests propòsits mitjançant mesures extretes d'una xarxa radar. En aquesta tesi, s'ha introduït el k-NN en un algoritme d'associació conjunt de rangs i Doppler per proporcionar una nova solució basada en quadrícules per distingir els blancs en escenaris multi-target (blancs) utilitzant una xarxa de radar OFDM completament multiestàtica. També s'ha desenvolupat un algoritme de fusió de dades per presentar estimacions de posicionament amb k-NN i estimacions de velocitat amb una expressió matemàtica de llaç tancat. Els resultats mostren un error comprensible d'associació i fusió, depenent de la resolució de la graella definida i del paràmetre k utilitzat. A més a més, la gran adaptabilitat del k-NN a totes les diferents hipòtesis a considerar en l'associació de dades fa que la tècnica no necessiti filtrar dades utilitzant algoritmes costosos en temps i en complexitat computacional. En el futur, aquest treball avançaria cap a la fusió de dades dinàmica, utilitzant algoritmes de seguiment d'objectes combinats amb l'associació de dades desenvolupada amb k-NN, amb l'objectiu d'oferir resultats altament fiables en el seguiment de blancs com ara drons.

La cirurgia de distracció del maxil·lar superior s’ha descrit per diversos autors com el procediment estàndard per el tractament de la hipoplàsia maxil·lar servera en pacients amb fissura llabipalatina. Altres autors han identificat els beneficis de l’ús de la planificació quirúrgica en 3D com una tècnica que millora el resultat final de la cirurgia i redueix els temps quirúrgics. Tot i així, a la bibliografia, existeixen molt pocs articles que descriguin la semblança entre la simulació del resultat obtinguda a la planificació quirúrgica i el resultat final del procediment quirúrgic. Així doncs, l’objectiu principal d’aquest treball és comparar el resultat postquirúrgic en 8 pacients sotmesos a cirurgia de distracció osteogènica del maxil·lar superior amb el resultat definit com a ideal durant el procediment de planificació quirúrgica en 3D. Per la realització del treball s’han analitzat 8 pacients amb fissura llabipalatina que presentaven una hipoplàsia maxil·lar severa, els quals van ser sotmesos a una cirurgíua de distracció osteogènica del maxil·lar superior amb prèvia realització de planificació quirúrgica en 3D. En tots els casos, durant la realització de la planificació quirúrgica es van dissenyar guies de tall i guies de posicionament que van permetre el posterior trasllat de la planificació quirúrgica al quiròfan. Un cop finalitzada la distracció i el temps de consolidació del moviment ossi, es va realitzar una sèrie d’imatges CT postquirúrgiques prèvies a la retirada dels distractors. Un cop finalitzat el procediment quirúrgic, s’ha obtingut una seqüència d’imatges de tomografia compuretitzada (TC), abans de retirar dels distractors. Aquestes imatges s’han segmentat i se’n ha extret els models poligonals que representen les superfícies òssies postquirúrgiques. S’han definit diverses variables, com el posicionament dels distractors, el posicionament dels claus d’osteosíntesis, la posició final dels incisius del maxil·lar i la posició final de la superfície