.: SIBGRAPI 2012 - Invited Speakers :.
Roberto Marcondes Cesar Jr
University of São Paulo, Institute of Mathematics and Statistics, Brazil
Title: Graph-based Pattern Recognition and Applications
Abstract: Structural pattern recognition plays a central role in many applications. Recent advances include new theoretical results, methods and successful applications. In the present talk, some recent graph-based methods for image segmentation will be shown. The presented methods include a new representation for graph-matching-based interactive segmentation and models for the analysis of spatial relations between objects. Applications will be presented and discussed.
Short Bio - Roberto Marcondes Cesar Jr é Professor Titular da Universidade de São Paulo (USP) desde 2008. Possui graduação em Ciência da Computação pela Universidade Estadual Paulista Júlio de Mesquita Filho (IBILCE - UNESP - 1992), mestrado em Engenharia Elétrica pela Universidade Estadual de Campinas (UNICAMP - 1993) e doutorado em Física pela Universidade de São Paulo (USP - 1997). Atualmente é Professor Titular do Departamento de Ciência da Computação - IME - USP. É membro da Coordenação de Área de Computação da FAPESP (2006-) e do Comitê de Avaliação Trienal da Capes (Computação (2007-2009; 2010-2012). É Diretor do Núcleo de Pesquisa em Bioinformática da USP e Chefe do Departamento de Ciência da Computação - IME-USP. Atua como J.A. do CTBE, Campinas. Tem experiência nas áreas de Ciência da Computação (modelagem matemática e engenharia), com ênfase em Processamento Gráfico (Graphics), atuando principalmente nos seguintes temas: visão computacional, reconhecimento de padrões; processamento de imagens e bioinformática.
University of Modena and Reggio Emilia, Italy
Title: People Detecting, Identifying and Searching in Video Surveillance and Multimedia Forensics
Abstract: Many results of pattern recognition research can be directly suitable for applications in video surveillance and multimedia forensics, concerning people security. The talk will present how some state-of-the art solutions, in classification, clustering, similarity search are exploitable for analyze the presence, the activity and the identity of persons in video. Recent advances in efficient search based on ''?particle'' windows? for detecting people shapes in image and video will be discussed. Then a short review of the state of the art of approaches for people re-identification in video footage will be presented, without covering aspects related with biometry, but underlying methods for assessing similarity of people appearance with 2D and 3D body models. Many possible applications for tracking people in cluttered and crowded context, searching for aspect similarities in real-time video surveillance and in off-line forensics will be illustrated.
Short Bio - Rita Cucchiara is Full Professor in Computer Engineering at the Faculty of Engineering, University of Modena. Her current interests include pattern recognition and computer vision for video surveillance, medical imaging, and multimedia. In multimedia she worked in models for semantic transcoding of videos. Now, she is involved in researches on video annotation and MPEG-7 video access. Medical imaging researches are oriented to colour analysis for skin lesion and melanoma classification. Video surveillance activity is devoted to new models of object segmentation, shadow detection, tracking, and people behaviour analysis indoor and outdoor applications. Eventually, these models are applied in intelligent transportation systems for traffic control. She is responsible of many Italian and International projects. Currently, she is responsible of a Research Unit in the “ European Network of Excellence on Digital Library” DELOS EU VI FP 2004-2008, and in Italian PRIN (Progetti di Rilevante Interesse Nazionale) for skin lesion analysis (2004-2006); she is scientific coordinator of the Regional Project LAICA (Laboratorio di Ambient Intelligent per una citta’ amica) with the Region Emilia Romagna and manages many projects with private companies (e.g. Traficon Belgium, world leader in traffic surveillance). Moreover, she is co-responsible with Prof. Piccardi of University of Technology of Sidney of a large project of Surveillance funded by Australian Ministry LP0668325- “Automatic real-time detection of infiltrated objects for security of airports and train stations”. She is author of more than 100 papers in national and international journals, and conference proceedings. She currently serve as reviewer for many international journals (e.g. IEEE Trans. on PAMI, IEEE Trans. on Circuit and Systems, Trans. on SMC, Trans. on Vehicular Technology, Trans. on Medical Imaging, Image and Vision Computing,..). She participated at scientific committees of international conferences (CVPR, ICME, ICPR, …) and symposia and organized special tracks in image processing and multimedia. She is member of GIRPR (Italy- associated with IAPR), AixIA (Ital. Assoc. Of Artificial Intelligence), ACM and IEEE Computer Society. In 2000 she organized the bi-annual School on Machine Vision of IAPR-IC. In 2005 she co-organized the 3rd Workshop of VSSN05 jointed with ACM Multimedia Conference and is currently editor of a special issue on Multimedia System Journal of ACM. In 2006 she is co-chair of the European School of Multimedia Digital Library. In 2007 she will be general chair of the 14th International Conference of Image Analysis and Processing, sponsored by IAPR.
INRIA Nancy Grand-Est/Loria, France
Title: Texture Synthesis for Large Virtual Environments
Abstract: A major difficulty in building vast, rich virtual worlds is the large amount of content that is required, in particular 3d meshes and textures. It is therefore not surprising that researchers in computer graphics started to consider methods to automatically generate this content. Textures in particular have received much attention: Images of materials tend to occupy a large amount of memory and their content is often of a stochastic nature. On the one hand this randomness makes them hard to compress efficiently as images, but on the other hand one can hope to describe their content through algorithms. However, being able to compute textures of a given appearance only partially solves the problem: The generated data must also be available in a form allowing for high quality, interactive rendering. In this talk I will describe the major challenges we are facing when synthesizing content for virtual worlds. I will focus on textures, describing the main families of approaches for their synthesis. I will also discuss which data structures may be used to hold the data being continuously generated as the user explores the scene. Along the talk I will point out directions of future research in this area.
Short Bio - Sylvain Lefebvre is a researcher at INRIA, France. He completed his PhD in 2004, on the topic of texturing and procedural texture generation using GPUs. He joined Microsoft Research (Redmond, USA) as a postdoctoral researcher during the year 2005, where he worked with Hugues Hoppe on runtime parallel texture synthesis. Their synthesizer was the first to run entirely on the GPU, allowing for direct interaction with the synthesized patterns. In 2006 he was recruited by INRIA as a permanent researcher. He is currently a member of the team ALICE in the Nancy Grand-Est INRIA center. His main research focuses are in procedural content generation -- often starting from examples -- end-user content manipulation and compact GPU-friendly data structures for interactive applications and games. Sylvain contributed several data-structures for parameterization-free texturing, such as spatial hashes that can be efficiently accessed from the GPU. Sylvain is an active member of the computer graphics research community, serving on the paper’s committee of major conferences such as SIGGRAPH and EUROGRAPHICS. He was short-paper co-chair for EUROGRAPHICS 2011. In 2010, he was honored with the EUROGRAPHICS young researcher award for his work on texture synthesis and GPU data-structures. In his vanishing spare time, Sylvain is an avid gamer and hobbyist game developer.
Université Paris-Est, Laboratoire d'Informatique Gaspard-Monge, Equipe A3SI, ESIEE Paris, France
Title: Morphological Filtering in Shape Spaces: Applications using Tree-Based Image Representations
Abstract: Connected operators are filtering tools that act by merging elementary regions of an image. A popular strategy is based on tree-based image representations: for example, one can compute a shape-based attribute on each node of the tree and keep only the nodes for which the attribute is sufficiently strong. This operation can be seen as a thresholding of the tree, seen as a graph whose nodes are weighted by the attribute. Rather than being satisfied with a mere thresholding, we propose to expand on this idea, and to apply connected filters on this latest graph. Consequently, the filtering is done not in the space of the image, but on the space of shapes build from the image. Such a processing is a generalization of the existing tree-based connected operators. Indeed, the framework includes classical existing connected operators by attributes. It also allows us to propose a class of novel connected operators from the leveling family, based on shape attributes. Finally, we also propose a novel class of self-dual connected operators that we call morphological shapings. Illustrations of the interest of the proposed approach will be provided on several tree-based image representations, including min-tree, max-tree, tree of shapes and hierarchical segmentations.
Short Bio - Laurent Najman received the Habilitation à Diriger les Recherches in 2006 from the University of Marne-la-Vallée, a Ph.D. of applied mathematics from Paris-Dauphine University in 1994 with the highest honour (Félicitations du Jury) and an “Ingénieur” degree from the Ecole des Mines de Paris in 1991. After earning his engineering degree, he worked in the Central Research Laboratories of Thomson-CSF for three years, working on some problems of infrared image segmentation using mathematical morphology. He then joined a start-up company named Animation Science in 1995, as director of research and development. The technology of particle systems for computer graphics and scientific visualisation, developed by the company under his technical leadership received several awards, including the “European Information Technology Prize 1997” awarded by the European Commission (Esprit programme) and by the European Council for Applied Science and Engineering and the “Hottest Products of the Year 1996” awarded by the Computer Graphics World journal. In 1998, he joined OCÉ Print Logic Technologies, as senior scientist. He worked there on various problem of image analysis dedicated to scanning and printing. In 2002, he joined the Informatics Department of ESIEE, Paris, where he is professor and a member of the Laboratoire d’Informatique Gaspard Monge, Université Paris-Est Marne-la-Vallée. His current research interest is discrete mathematical morphology and discrete optimization.