Rodrigo Luis de Souza da Silva

Interactive mesh generation with local deformations in multiresolution

This work presents a method to model a spherical mesh by modifying its heightmap in an augmented reality environment. Our contribution is the use of the hierarchical structure of semiregular A4-8 meshes to represent a dynamic deformable mesh suitable for modeling. It defines only a fraction of the overall terrain that is subjected to local deformations. The modeling of spherical terrains is achieved with proper subdivision constraints at the singularities of the parametric space. An error metric dependent on the observer and on the geometry of the topography was used to provide fast visualization and editing. The results demonstrate that the use of the A4-8 mesh combined with the tangible augmented reality system is flexible to shape spherical terrains and can be easily modified to deal with other topologies, such as the torus and the cylinder.

Ferramenta de Apoio ao ensino de Física utilizando Realidade Aumentada

Resumo Este artigo apresenta uma ferramenta de apoio ao ensino de fisica utilizando Realidade Aumentada. Os principais conceitos sobre Realidade Aumentada e os frameworks da area serao apresentados. As funcoes que simulam os eventos fisicos serao explicadas e desenvolvidas. Neste trabalho nenhuma biblioteca classica de simulacao fisica foi utilizada. A ferramenta foi dividida em varios modulos da fisica e, para cada um destes modulos, um sistema baseado em Realidade Aumentada foi proposto. A proposta central deste trabalho e mostrar que ferramentas baseadas em Realidade Aumentada podem potencialmente tornar o estudo de fisica mais ludico e interessante.

A Method to Automate the Calibration of a Multiple Fiducial Marker Setup

This work proposes a new method to automate the calibration of multiple fiducial markers distributed over an Augmented Reality scene, in order to obtain stable relations, mainly among markers that will never be visible in a same video frame. The main feature of the method is to maintain all relationships between markers in each video frame. These relationships will be updated during calibration in order to obtain an accurate result. During the execution, the referential (base) marker of the system may change when markers with smaller error appears in the scene. The application produces the output configuration file of a multiple marker augmented reality environment in the same format used by AVRLib and ARToolKit libraries.

Virtual Reality as a Support Tool for the Treatment of People with Intellectual and Multiple Disabilities: A Systematic Literature Review

Since the emergence of virtual reality (VR) technologies, many researchers have argued on the benefits of their use for people with intellectual and multiple disabilities. However, up to this date there is not a single study that presents a detailed overview of the state of the art in virtual reality as a support tool for the treatment of people with intellectual and multiple disabilities, as well as Autism and Down Syndrome. The aim of this study is to provide a detailed overview of the state of the art in the virtual reality area focusing on people with multiple disabilities, that encompasses intellectual and physical disabilities. There is still no consensus on the effectiveness of VR-based treatments. Virtual reality can offer rich environment and features, but most of the researches focuses only in the experience to be inside a virtual place without taking advantage of what benefits VR provide us. Furthermore, most of our selected studies used non-immersive VR and AR. Thus, immersive VR is an open field with many opportunities to be explored. We believe VR has great potential to be effective in the treatment of people with intellectual and multiple disabilities.

Uma Arquitetura de Software para Bibliotecas de Realidade Aumentada Baseada em Marcadores Naturais

Um importante topico de estudo no campo da Realidade Aumentada e o desenvolvimento de sistemas que nao necessitam de marcadores fiduciais. Nesses sis- temas sao utilizadas somente caracteristicas naturais da cena para posicionar a câmera e realizar a projecao. Essa caracteristica torna as aplicacoes mais inteligiveis e sua utilizacao e imediata. Entretanto, nao existem muitas bibliotecas que auxiliem o de- senvolvimento de aplicacoes em Markerless Augmented Reality (MAR) e, as poucas que existem, geralmente sao de codigo fechado. Este artigo apresenta uma arquitetura de software que pode ser utilizada para a criacao de bibliotecas de MAR. Todos os recursos de Visao Computacional serao oriundos da biblioteca OpenCV.

Gesture-Based Manipulation of Virtual Terrains on an Augmented Reality Environment

This paper presents a free hand interface for modeling virtual terrains in an augmented reality environment. A secondary contribution of this work is to present a study of a set of suitable gestures that allows the manipulation of virtual terrains interactively, using augmented reality markers and the Leap Motion Controller. To demonstrate the study, an application was developed to allow the user to interact with the virtual terrain directly with his/her bare hands. Virtual objects were augmented using fiducial markers and the detection is done through the AVRLib Library. In our test layout, both markers and controller are co-planar. The application was tested and evaluated by six subjects. None of the subjects had previous knowledge on how to use the Leap Motion controller or the nature of the application.

A Video Self-descriptor Based on Sparse Trajectory Clustering

In order to describe the main movement of the video a new motion descriptor is proposed in this work. We combine two methods for estimating the motion between frames: block matching and brightness gradient of image. In this work we use a variable size block matching algorithm to extract displacement vectors as a motion information. The cross product between the block matching vector and the gradient is used to obtain the displacement vectors. These vectors are computed in a frame sequence, obtaining the block trajectory which contains the temporal information. The block matching vectors are also used to cluster the sparse trajectories according to their shape. The proposed method computes this information to obtain orientation tensors and to generate the final descriptor. The global tensor descriptor is evaluated by classification of KTH, UCF11 and Hollywood2 video datasets with a non-linear SVM classifier. Results indicate that our sparse trajectories method is competitive in comparison to the well known dense trajectories approach, using orientation tensors, besides requiring less computational effort.

A Video Descriptor Using Orientation Tensors and Shape-Based Trajectory Clustering

Dense trajectories have been shown as a very promising method in the human action recognition field. In this paper, we propose a new kind of video descriptor, generated from the relationship between the trajectory’s optical flow with the gradient field in its neighborhood. Orientation tensors are used to accumulate relevant information over the video, representing the tendency of direction in the descriptor space for that kind of movement. Furthermore, a method to cluster trajectories using their shape is proposed. This method allows us to accumulate different motion patterns in different tensors and easier distinguish trajectories that are created by real movements from the trajectories created by the camera’s movement. The proposed method is capable to achieve the best known recognition rates for methods based on the self-descriptor constraint in popular datasets — Hollywood2 (up to 46%) and KTH (up to 94%).

Iterative Remeshing for Edge Length Interval Constraining

This paper presents an iterative method to remesh an arbitrary surface into a mesh with all edge lengths within an interval. The process starts with a triangular 2-manifold mesh. It uses stellar operations to achieve the necessary amount of vertices and triangles. Subsequently, it applies a constrained version of the Laplacian filter in order achieve a more uniform distribution of the vertices over the surface. In order to prevent the natural shrink caused by the Laplacian filter, we perform a projection over the original surface. We also apply a post processing step to correct the lengths of troubling edges. Our method results in a regular mesh, with vertices uniformly distributed. The dual mesh obtained can be useful for several applications. The main contribution of this work is a new approach for edge length equalization, with explicit constraints definition, lower global geometry losses and lower memory cost if compared to previous works.

Analysis of a high definition camera-projector video system for geometry reconstruction

The present work describes the development of a high definition camera-projector system (with resolutions up to 1920x1080 and 1280x720). The steps and processes that lead to the reconstruction of geometry, from system setup to triangulation, are detailed. A structured light coding scheme that offers a flexible number of stripes for projection was used. One of the objectives of this work is to estimate the limit number of stripes possible within the current resolutions of high definition video. This limit number is the one that provides dense geometry reconstruction with low accuracy and precision errors. To evaluate the reconstructions, we propose a protocol for error measurement. The geometry of general objects is also presented for qualitative evaluation. Our results show that low cost but effective 3D scanners can be built with high definition video devices providing compressed data.