SIBGRAPI 2007 Technical Poster Session
Preface
Program Committee
Reviewers
Authors
Exhibition Sessions: The posters will be exposed in the Coffee-break room from 9:00AM until 4:30PM either in Monday or in Tuesday. To avoid confusions, we recommend authors to hang them until 9:00AM. One of its author is supposed to be close to it during the coffee-break for providing more detailed explanations. Every participant may vote for the best presentation. The award selection is based on the voting records and on the reviewer's rankings. Image Processing Computer Graphics, Vision and Modeling
Download all posters
Caderno de Pôsters em pdf
Monday, October 8, 2007
10:00AM - 10:30AM and 4:00PM - 4:30PM	Image Processing - Chair:
Coffee-Break	Frontal Sinus Recognition Using Image Foresting Transform and Shape Context (pp 31 - 32)
Room	Juan Falguera, Sao Paulo State University, Brazil Aparecido Marana, UNESP, Brazil Fernanda Falguera, Unesp, Brazil
	Abstract. It has been established in prior investigations that the radiographic pattern of frontal sinus is highly variable and unique for every individual. This paper presents a frontal sinus recognition method, for human identification, based on DIFT and Shape Context algorithms. Experimental results show the effectiveness of the proposed method.
	Super-Resolution Image Reconstruction using the Discontinuity Adaptive ICM (pp 17 - 18)
	Murillo Homem, Ana Luísa Martins, Universidade Federal de São Carlos, Brazil Nelson Mascarenhas, Universidade Federal de São Carlos - UFSCar, Brazil
	Abstract. We propose a Bayesian approach for the super resolution image reconstruction (SRIR) problem using a Markov random field (MRF) for image characterization. SRIR consists in using a set of low-resolution (LR) images from the same scene to generate a high-resolution (HR) estimate of the original object. Using a Bayesian formulation, it is possible to incorporate previously known spatial information about the HR image to be estimated. In our approach, the iterated conditional modes (ICM) algorithm is used to find the maximum a posteriori (MAP) solution, and a discontinuity adaptive framework is used to overcome the oversmoothness inherent to MAP-MRF formulations. To evaluate the capability of the algorithm in reconstructing the actual image, we used the universal image quality index (UIQI). According to this index, the proposed method produced accurate results.
	Estimation of Non-Homogeneous Potts-Strauss MRF Model Parameters on Higher-Order Neighborhood Systems by Maximum Pseudo-Likelihood (pp 3 - 4)
	Alexandre Levada, Universidade de São Paulo, Brazil Nelson Mascarenhas, Universidade Federal de São Carlos - UFSCar, Brazil
	Abstract. This paper addresses the problem of maximum pseudo-likelihood estimation of the non-homogeneous Potts-Strauss image model parameters using higher- order non-causal neighborhood systems in a computationally efficient way. The motivation is the development of a new methodology for contextual classification that uses combination of sub-optimal MRF algorithms for multispectral image classification, which requires accurate parameters estimation. The results show that the method is consistent with real image data and in the presence of random noise.
	Comparison of Three Different Derivative Approaches Aiming at Estimation of Image Movement (pp 41 - 42)
	Mauricio Higa, University of Sao Paulo, Brazil Marina Rebelo, Heart Institute - University of Sao Paulo Medical School, Brazil Carlos Santos, University of São Paulo, Brazil Marco Gutierrez, Heart Institute, University of Sao Paulo Medical School, Brazil
	Abstract. The use of optical flow techniques to extract the velocity from the cardiac movement has to take into account the computational inaccuracy raised from the derivative operator over discrete data. This study presents a comparison of three different derivative approaches (two based on linear and other based on non-linear filtering) to find out the best solution. Results of the experiments are compared using a structural distortion based image quality metric.
	A Statistical Approach for Image Interpolation (pp 61 - 62)
	Murillo Homem, Universidade Federal de São Carlos, Brazil Nelson Mascarenhas, Universidade Federal de São Carlos - UFSCar, Brazil
	Abstract. Interpolation is an image processing operation for improve the resolution of a digital image. In this work, following the orthogonality principle, and under the assumption that the actual image is a locally stationary random process, we propose an alternative scheme for image interpolation. In our approach, the computational complexity is similar to the first-order spline algorithm. The algorithm was compared with classical B-spline methods and also with a statistical interpolator previously proposed in the literature. According to the normalized mean square error criteria, the proposed method produced accurate results.
	Cerebral tissue segmentation using q-entropy in multiple sclerosis magnetic resonance images (pp 21 - 22)
	Paula Diniz, USP, Brazil Antônio Carlos dos Santos, University of São Paulo, Brazil Luiz Otavio Murta Junior, FFCLRP-USP, Brazil
	Abstract. The brain volume measurements have important clinical applications in the treatment of neurodegenerative illnesses. The segmentation and the volume calculation are important in the medical context to provide information to assist the physician in disease diagnostic and prognostic. Furthermore, it can improve the speed of the diagnostic. The objective of this article is described the development and evaluation of a tool for calculation of brain volume using Tsallis Entropy.
	Application of image restoration algorithms in vibro-acoustography images (pp 29 - 30)
	Talita Perciano, Universidade Federal de São Carlos, Brazil Nelson Mascarenhas, Universidade Federal de São Carlos - UFSCar, Brazil Alejandro Frery, Glauber T. Silva, Universidade Federal de Alagoas, Brazil
	Abstract. Vibro-acoustography (VA) is an imaging modality that produces an image of the mechanical response of an object to a localized dynamic radiation force of an ultrasound field. This technique has been studied and used in clinical applications as to image calcification in breast tissue and arteries. This paper presents the application of restoration algorithms to VA images.
	Ilumination Normalization Methods for Face Recognition (pp 37 - 38)
	Michelle Magalhães Mendonça, Universidade de São Paulo, Brazil Juliana Denipote, Ricardo Augusto Souza Fernandes, University of São Paulo, Brazil Maria Stela Veludo de Paiva, EESC-Universidade de São Paulo, Brazil
	Abstract. This paper describes three methods (logAbout, homomorfic filter and wavelet) for illumination normalization and compares their contributions for posterior face recognition using Principal Component Analysis (PCA).
	Analyzing Polarimetric Imagery with G0p Mixture Models and SEM Algorithm (pp 71 - 72)
	Michelle Horta, University of São Paulo-São Carlos Campus, Brazil Nelson Mascarenhas, Universidade Federal de São Carlos - UFSCar, Brazil Alejandro Frery, Universidade Federal de Alagoas, Brazil
	Abstract. This paper presents the use of a finite mixture model for multi-look polarimetric SAR image analysis. The pixels are complex covariance matrices set as a G0p mixture distribution. The parameters are estimated with the SEM algorithm. Experimental results on real SAR data are reported, showing that a careful statistical model is important.
	Cardiac Phase Detection in Intravascular Ultrasound Images (pp 33 - 34)
	Monica Matsumoto, InCor, FMUSP, Brazil Takashi Yoneyama, Instituto Tecnológico de Aeronautica, ITA-IEE-IEES, Brazil Sérgio Furuie, InCor, HC-FMUSP, Brazil
	Abstract. Image gating is a problem related to image modalities which involve quasi-periodic moving organs, such as the heart. Therefore, during intravascular ultrasound (IVUS) examination, with automatic pullback, there is cardiac movement interference. This work aims to obtain gated images based on the images themselves, so it would be possible to reconstruct 3D coronaries with temporal and spatial accuracy. From the images, we extracted signals based on average intensity (AI), and, from consecutive frames, average intensity difference (AID), cross-correlation coefficient (CC) and a new approach using mutual information (MI). Firstly, our method was tested in simulated images, with different speckle noise level, contrast and patient''s characteristics. The results have shown our method was able to achieve more than 95.0% of true positives and 2.3% of false positives, for all signals. Afterwards, we tested in a real IVUS examination, with ECG as gold-standard, where there were achieved 97.4% of true positives (CC and MI), and 2.5% of false positives.
	Total Cholesterol determination Using Digital Image Processing (pp 19 - 20)
	Leandro Luís Galdino de Oliveira, Eduardo Simões Albuquerque, Paula Lorenzoni Paste de Oliveira, Paulo Luiz Carvalho Francescantonio, Universidade Católica de Goiás, Brazil
	Abstract. We present a low cost Digital Image Processing method to measure cholesterol levels. Images are captured from an ELISA plate using a regular digital camera. The results showed strong correlation between the analysis performed with a spectrophotometer and the proposed method. The method can reduce cost and provide access to the exams in areas distant from urban areas.
	Improved Inference for the GA0 Distribution (pp 67 - 68)
	Michel da Silva, Universidade Federal de Minas Gerais, Brazil Francisco Cribari-Neto, Universidade Federal de Pernambuco, Brazil Alejandro Frery, Universidade Federal de Alagoas, Brazil
	Abstract. This paper presents adjusted profile likelihoods for alpha, the roughness parameter of GA0(alpha, gamma, L) distribution. This distribution has been widely used in the modelling of data corrupted by speckle noise (SAR images). We focus on point estimation and on signalized likelihood ratio tests. As far as point estimation is concerned, the numerical evidence presented in the paper favors the Cox and Reid''s adjustment (1987), and in what concerns signalized likelihood ratio tests, the results favor the approximation to Barndorff-Nielsen''s adjustment based on the results in Fraser and Reid (1995). An application to real synthetic aperture radar imagery is presented.
	Automatic Texture Segmentation Based on k-means Clustering and Co-occurrence Features (pp 43 - 44)
	Lucas Bastos, Universidade Federal Fluminense, Brazil Aura Conci, UFF, Brazil
	Abstract. This work presents a method for automatic texture segmentation based on k-means clustering technique and co-occurrence texture features. A set of features was extracted from 256 grey-level co-occurrence information. These features were used to segment image regions regarding the textural homogeneity of its areas. As the process of calculating co-occurrence information demands the majority of computational time required,we propose a new methodology based on a grey-level co-occurrence indexed list (GLCIL) for fast element access, highly optimizing this step in the algorithm. Besides that, we compare the efficiency of the proposed method against other well known algorithms. The experiments show that GLCIL is the most efficient method in terms of computational time. Additionally, traditional Brodatz textures and other literature examples were tested to evaluate the appropriateness and robustness of the method.
	A New Approach for Creating Polar Maps of Three-Dimensional Cardiac Perfusion Images (pp 53 - 54)
	Lucas Ferrari de Oliveira, Bruno Zanchet, Rodrigo Barros, Universidade Federal de Pelotas, Brazil Marcus Simões, Faculdade de Medicina de Ribeirão Preto, Brazil
	Abstract. We present a new approach for creating polar maps of three-dimensional cardiac perfusion images. A polar map is a two-dimensional plot of the reconstructed volume from the SPECT study, which provides concise information on myocardial perfusion in a single easily interpretable image. Unlike the traditional approaches, our algorithm is automatic. It relies on image registration techniques to align the patient exam to a previously computed model, avoiding the need of manual intervention by a specialist. Preliminary analysis made by a specialist in nuclear medicine indicates that this new approach is comparable to the gold standard algorithm
	Images of Cutaneous Ulcers Classified by Artificial Neural Networks (pp 9 - 10)
	ANDRÉ Tarallo, UNIVERSITY OF SÃO PAULO, Brazil Adilson Gonzaga, Marco Andrey Frade, Wellington Gouveia, University of São Paulo, Brazil
	Abstract. Treatments of leg ulcers are generally made by direct manipulation for analysis of its evolution. The treatment efficiency is observed through the reduction of the size of ulcers in relation to the amount of tissues found in their beds, which are classified as granulated/slough. These results are usually obtained through analyses performed after consultation due to the time these analyses take. This work proposes a new non-invasive technique for the follow-up of treatments aimed at cutaneous ulcers. In this technique, it was proposed that digital photos of cutaneous ulcers would be submitted to an artificial neural network, so that all surrounding the wound except for the wound itself could be extracted (skin/background), thus obtaining the ulcerated area. Computer vision techniques have been applied in order to classify the different types of tissues in the ulcer bed. The results obtained have been compared with the results obtained by Image J software.
	Quantitative Analysis of SPECT Myocardial Perfusion and Assessment of Myocardium Defect Regions through Image Processing Techniques (pp 51 - 52)
	Rodrigo Barros, Lucas Ferrari de Oliveira, Universidade Federal de Pelotas, Brazil Marcus Simões, Faculdade de Medicina de Ribeirão Preto, Brazil
	Abstract. In this paper we present a new approach for quantitative analysis of myocardial perfusion SPECT studies and assessment of myocardium defect regions. Myocardial perfusion SPECT is widely used in the evaluation of post-infarction patients, in order to foretell the patients future conditions. Our approach is based on the use of polar maps, a two-dimensional plot of the reconstructed volume from the SPECT study. Using an automatic technique for polar map creation, based on image registration techniques, we were able to generate six different polar maps, allowing medical visual interpretation and analysis of myocardium defect regions.
	A Probabilistic Approach to Skin Detection (pp 39 - 40)
	Fernando Cardoso, Herman Gomes, Universidade Federal de Campina Grande, Brazil
	Abstract. In this paper, we present an approach to skin detection in images, which labels each pixel as skin or non-skin based on a simple probability model acquired from a database of mannualy labeled images. We tested several variations of this approach (using different color spaces and pre-processing techniques) and the achieved accuracy seemed quite encouraging with a detection rate of 77.79%.
	Natural Computing Techniques for Data Clustering and Image Segmentation (pp 63 - 64)
	Jackson Souza, Universidade Federal do Rio Grande do Norte, Brazil José Alfredo F. Costa, UFRN, Brazil
	Abstract. This paper presents innovative ways to solve data clustering and image segmentation using Natural computing, a novel approach to solve real life problems inspired in the life. Evolutionary Computing, which is based on the concepts of the evolutionary biology and individual-to-population adaptation, and Swarm Intelligence, which is inspired in the behavior of individuals that, in group, try to achieve better results for a complex optimization problem, are detailed and very experimental results present a comparison between algorithms'' implementations.
	Content based image retrieval using tree-structured self-organizing maps (pp 69 - 70)
	José Alfredo F. Costa, UFRN, Brazil Raquel Esperanza Patiño Escarcina, Universidade Federal do Rio Grande do Norte, Brazil
	Abstract. Content-based image retrieval systems are designed to proivide effective access to image databases, based on their visual contents and according to a given criteria. This paper focuses the image seaching based on descriptors automatically extracted from the images. It is presented a scheme that decomposes the image collection in a hierarchy of clusters using tree-structured self-organizing maps. The present approach uses different sets of features in each level of the hierarchy. Subsets of images are clustered in each level of the tree of maps, associated to regions of neurons, and are used to train subsequent maps in the hierarchy. The tree-structured self-organizing map algorithm may be seen as a recursive image data partition clustering method. Using of this architecture is motivated not only by the reduction of complexity but also as a weak resemblance of the visual information system that process different types of information in different layers of neurons.
Tuesday, October 9, 2007
10:00AM - 10:30AM and 4:00PM - 4:30PM	Computer Graphics, Vision and Modeling - Chair:
Coffee-Break	On the Simulation of Ocean Waves in Real-Time Using the GPU (pp 15 - 16)
Room	Alex Salgado, Fluminense Federal University, Brazil Aura Conci, UFF, Brazil
	Abstract. Nowadays, due to GPU processing power, it is possible to use advanced rendering techniques with great realism in real-time applications. This work simulates the ocean wave behavior processing all geometric computation and rendering in GPU. The shape is defined by the Gerstner´s equation considering the movement of each particles of water. The deep sea topology is considered. It is possible to tackle the representation in deep water as well as in shallow water. The ocean floor profoundness and inclination influence on the shape and form of the particle displacement are considered. The model represents also the wave refraction and the wind or gale effect on the wave. This novel approach simulates the breaking waves near the shore. The real-time rendering technique implemented in this work uses combinations of advanced tangent-space reflective bump mapping and environment mapping plus Fresnel reflection and HDR.
	A Simplified Approach for Animation of Deformable Objects (pp 47 - 49)
	Guina Sotomayor Alzamora, Federal University of Rio de Janeiro, Brazil Yalmar Atencio, COPPE/UFRJ, Brazil Claudio Esperança, UFRJ, Brazil
	Abstract. We present a simplified approach for animation of geometrically complex deformable objects represented as tetrahedral meshes. Our prototype system detects and responds to collisions of objects subject to elastic deformations of variable stiffness. The proposed approach combines several techniques, namely, collision detection using a spatial hashed grid [6], consistent penetration depth using propagation [1], contact surface computed according to [5] and projection [2], the animation is based in a modal analysis scheme using an explicit-implicit integrator as in [3]. Preliminary results show that collisions between objects containing several thousand tetrahedra can be animated in real-time.
	Muan: Animation for the rest of us (pp 13 - 14)
	Margareth Varela, Instituto Nacional de Matematica Pura e Aplicada, Brazil Luiz Velho, IMPA, Brazil Hedlena Maria Bezerra, Pontificia Universidade Catolica do Rio de Janeiro, Brazil Bruno Madeira, IMPA, Brazil Marcos Magalhaes, Anima Mundi, Brazil
	Abstract. MUAN is a Stop Motion Animation System. The MUAN system allows animation movie production offering an integrated kit of hardware and software. It consists of a computer with Linux, a video camera and support over necessary accessories for interconnection. The software was developed by the team of VISGRAF Laboratory (IMPA), in cooperation with ANIMA MUNDI and IBM Brazil, and consists of programs to create, edit, manipulate and visualize animations.
	Augmented Reality for Life Support Training (pp 55 - 56)
	Fabrício Pretto, PUCRS, Brazil Isabel Manssour, PUCRS - Pontifícia Universidade Católica do Rio Grande do Sul, Brazil Marcio Pinho, Maria Helena Itaqui Lopes, Emerson Rodrigues da Silva, Pontifícia Universidade Católica do Rio Grande do Sul, Brazil
	Abstract. The area of Medical Qualification in Life Support training is being constantly improved. However, many problems are identified in the training process, such as the lack of realism in the exercises and the low student involvement. In order to qualify the learning process, the ARLIST project (Augmented Reality for Life Support Training) is being developed to add computational resources as sound and images, in the manikins used in the training courses. Through Augmented Reality techniques, the use of OpenGL and some computational resources (e.g. projector ad video camera), it is possible to build an application that defines the images and sounds that should appear in accordance to the patient clinical state. This ongoing work is being here described.
	Application of Computer Graphics to Headlights Dynamic Simulation Tests (pp 5 - 6)
	Alexandre Faria, UFMG Federal University of Minas Gerais, Brazil Arnaldo de Albuquerque, UFMG, Brazil
	Abstract. In this work, we introduce an application of computer graphics to the validation of an automotive illumination system in dynamics conditions. We propose the creation of a virtual road, where it is possible to drive the vehicle for different kinds of tracks and study the performance of the headlights in dynamics conditions on the road. The projection of the light is directly affected by the dynamics behavior of the vehicle and texture of the road.
	Soft Segmentation for Comparative Image Editing (pp 7 - 8)
	Anna Regina Corbo, Luiz Velho, IMPA, Brazil
	Abstract. We present a new tool for comparative image editing of HDR and LDR images that preserves edges by doing a soft segmentation. The user selects with strokes different regions to be modified and has control over all the regions simultaneously to make adjustments in parameters such as saturation, contrast and exposure. The final mask that maps those regions preserves the most important transformation areas and the smoothness of transitional regions.
	Shape-aware as rigid as possible deformation (pp 23 - 24)
	Alvaro Cuno, Universidade Federal do Rio de Janeiro, Brazil Claudio Esperança, UFRJ, Brazil Antonio Alberto Fernandes de Oliveira, Paulo Cavalcanti, Universidade Federal do Rio de Janeiro, Brazil
	Abstract. We propose a formulation capable of deforming meshes in a shape-sensible way. We explain how to adapt the original space-deforming algorithm [2] into a skeleton-driven deformation scheme, where more sensitivity to the mesh geometry is achieved. The more natural results make it possible to use the technique for character animation.
	Efficient viewshed computation on external memory DEM terrains (pp 45 - 46)
	Mirella Magalhaes, Salles Magalhaes, Universidade Federal de Vicosa, Brazil Marcus Andrade, University of Vicosa, Brazil
	Abstract. Many GIS applications require efficient algorithms to manipulate huge volume of data about terrains stored in external memory. One of these applications is the viewshed computation which consists in obtaining the points that can be viewed by a given point. In this paper, we present an efficient algorithm to compute the viewshed on huge terrains stored in external memory.
	Temporal-PEx: Similarity-based Visualization of Time Series (pp 35 - 36)
	Aretha Alencar, University of São Paulo, Brazil Maria Cristina de Oliveira, ICMC-USP/São Carlos, Brazil Fernando Paulovich, Universidade de São Paulo, Brazil Rosane Minghim, São Paulo University, Brazil Marinho Andrade, Universidade de São Paulo, Brazil
	Abstract. Time series analysis poses many challenges to professionals in a wide range of domains. Several visualization solutions integrated with mining algorithms have been proposed for exploratory tasks on time series collections. As the data sets grow large, though, the visual alternatives do not allow for a good association between similar time series. In this paper we introduce a visual representation or large time series data sets generated by multidimensional projections based on distance measures.
	Memory Organization for Invariant Object Recognition and Categorization (pp 11 - 12)
	Guillermo Donatti, Ruhr-Universität Bochum, Germany
	Abstract. The integration of bottom-up with top-down object processing has always been a topic of major concern in computer vision. However, while a lot is known about feature extraction, the knowledge-driven aspect of perception has been recognized as important, but hard to probe experimentally and difficult to implement in computer vision systems. How object knowledge must be organized so that it supports scene perception and can be acquired automatically is a research problem of outstanding significance for the biological, the psychological, and the computational approach to understand perception. The present work aims to develop an object memory model which can provide fast retrieval, and robust recognition and categorization. The underlying data structure of this model is inspired by the neural network structure of the human brain, connecting similar object views with excitatory synapses and using inhibitory synapses to separate different ones. The insights that derive from building such a computational theory and the properties of the resulting model have implications for strategies and experimental paradigms to analyze human object memory as well as technical applications for robotics and computer vision.
	Construction Of Georeferenced Mosaics Using Small Format Aerial Images (pp 65 - 66)
	Natal Cordeiro, Bruno Motta de Carvalho, Universidade Federal do Rio Grande do Norte, Brazil
	Abstract. We propose to use small format aerial images (SFAI) considered as not controled, and stereo-photogrammetry techniques for construction of georreferenced mosaics. The images are obtained using a simple digital camera coupled to a small, radio controled helicopter. Techniques for removing distortions are applied and the relative orientation of the models are computed using perspective geometry. Ground truth points are used for absolut orientation, and a definition of scale and a coordinate system relate image measures to the ground.
	A Computer-Aided System for Indexing People in Historical Images (pp 57 - 58)
	David Flam, Federal University of Minas Gerais, Brazil Camillo Jorge Santos Oliveira, Universidade Federal de Minas Gerais, Brazil Arnaldo de Albuquerque, UFMG, Brazil
	Abstract. Our work describes a computer aided system for indexing people in historical images based on facial recognition. A prototype has already been built and it is currently being applied to selected pictures from the Mineiro Public Archive´s digital photographic database. The architecture of the system is described and explained briefly. The methods used in our work for facial detection and identification are presented, along with some results. Finally, further directions and possible improvements are given.
	Detection of Generic Conic Form Parameters Using Hough Transform (pp 1 - 2)
	Maysa Macedo, Universidade Federal Fluminense, Brazil Aura Conci, UFF, Brazil
	Abstract. Hough Transform (HT) is a method for shape extraction that uses a parameter accumulator array. Based on the analysis of generic conic equation it is possible to establish a robust approach for conic shape identification in images if some aspects are respected. This paper introduces a unique methodology to detect any conic equation parameters using the HT idea. The basis of the formulation here presented is the use of polar coordinates to detect an open or closed form, and the parameters search sequence. In this way we can identify complex forms through of the union of several conic detections.
	Semi-Supervised Support Vector Rainfall Estimation Using Satellite Images (pp 27 - 28)
	Greice Martins de Freitas, Ana Ana Maria Heuminski de Avila, Meteorological and Climatic Research Center Applied to Agriculture, Brazil João Paulo Papa, State University of Campinas, Brazil
	Abstract. In this paper we introduce the use of semi-supervised support vector machines for rainfall estimation using images obtained from visible and infrared NOAA satellite channels. Two experiments were performed, one involving tradicional SVM and other using semi-supervised SVM. The semi-supervised support vector machines approach outperforms SVM in our experiments, with can be seen as a good methodology for rainfall satellite estimation, due to the large amount of unlabeled data.
	A Practical Approach on Human Authentication Using Biometrics (pp 59 - 60)
	Marcelo Fornazin, Faculdade de Cîências - Universidade Estadual Paulista "Júlio de Mesquita Filho" - Unesp, Brazil Marcos Cavenaghi, UNESP, Brazil Danilo Netto Jr., Universidade Estadual Paulista - Júlio de Mesquita Filho, Brazil
	Abstract. This is a work in progress, presenting concepts and techniques directly related to the practical use of the Fuzzy Vault Scheme (FVS) biometrics cryptosystem on human authentication using biometrics. Many applications handle confidential information from their users. These information need to be handled in a secure way. One way of handling information securely is through cryptography, where the use of passwords is a common practice. Passwords are the weakest link on a conventional cryptosystem, because passwords can be easily copied or stolen. Biometry is a way to overcome issues regarded the usage of passwords. Biometrics signals are more secure, but they also need to be protected. To protect biometrics signals one can use cryptography based on biometrics keys. But, due to the variability of the biometrics signal, biometrics cannot be directly used as cryptographic keys of a traditional cryptosystem. Cryptosystems that address to overcome these issues are called Biometrics Cryptosystems.
	itkFlowRun - a Visual Programming Tool for ITK Image Filters (pp 49 - 50)
	Diego dos Santos, State University of Campinas (UNICAMP), Brazil Eduardo Costa, State University of Campinas, Brazil Marco Gutierrez, Heart Institute, University of Sao Paulo Medical School, Brazil
	Abstract. itkFlowRun is an open source software that aims to be an extension of Insight Registration and Segmentation Toolkit (ITK) to rapidly and easily create image filters pipeline through a visual programming tool environment. It has been developed for Linux environment using the following libraries: Fast Light Toolkit (FLTK), Boost and ITK. This paper presents an overview of the main functionalities and describes the software architecture and plug-in mechanism, allowing third-party plug-ins to be developed. itkFlowRun is still being developed and will be released in the near future.
	3DbyStep: A Tool For Authoring 3D Presentations (pp 25 - 26)
	Elisabete Nogueira, Rio de Janeiro Federal University, Brazil Claudio Esperança, UFRJ, Brazil Victor Bursztyn, Federal University of Rio de Janeiro, Brazil
	Abstract. This work presents, 3DbyStep, an interactive authoring tool for developing 3D training contents. The application is able to combine text, pictures, sound and 3D models in an integrated environment. The produced contents is suitable for using as a standard presentation and also as a means to support self-study in the sense that it can be navigated interactively in a non-linear fashion. 3DbyStep is mainly aimed at application areas where the ability to view 3D models is crucial for conveying key information.