Einzusetzende Mittel - Intelligente Systeme
Transcrição
Einzusetzende Mittel - Intelligente Systeme
Seminar Vorbesprechung „Intelligente Systeme“, Sommersemester 2011 Betreut von: Prof. Dr. Pauli Dipl.-Inf. Bürger Dipl.-Inform. Herwig Dipl.-Inform. Hoefinghof Dipl.-Inform. Korn Seminar Intelligente Systeme 13.04.2012 1 Organisatorisches Dieser Vortrag sowie weitere aktuelle Informationen stehen unter http://www.is.uni-due.de/seminar zur Verfügung. Seminar Intelligente Systeme 13.04.2012 2 Organisatorisches Anforderungen: – Einarbeitung in das gewählte Thema – mehrmalige Besprechung mit dem jeweiligen Betreuer – Erstellen von Vortragsfolien im PDF oder PowerPointFormat (max 30 Folien) – Erstellen einer schriftlichen Ausarbeitung im PDF-Format (ca. 20 Seiten) – – Vorbereitung auf die Vorträge der anderen Teilnehmer Teilnahme an der Diskussion nach den Vorträgen Seminar Intelligente Systeme 13.04.2012 3 Organisatorisches Themenvergabe 1. Besprechung -7 Weitere Besprechungen vorläufige Folien -2 Abgabe der Folien -1 Vorbereitung Vortrag 0 Diskussion Abgabe der Ausarbeitung +3 t t in Wochen Seminar Intelligente Systeme 13.04.2012 4 Automatic Panorama Image Stitching Automatic Panorama Image Stitching Literatur: – – – Brown, Lowe; Automatic Panoramic Image Stitching using Invariant Features, International Journal of Computer Vision, 2007 Eden,Berkeley, Uyttendaele, Szeliski; Seamless Image Stitching of Scenes with Large Motions and Exposure Differences, Conference on Computer Vision and Pattern Recognition, 2006 Nomura, Zhang, Nayar; Scene Collages and Flexible Camera Arrays, Eurographics Symposium on Rendering (2007) Einzusetzende Mittel: Robuste Bildmerkmale, Optimierung Visual Attention for Rapid Scene Analysis Hierarchie „interessanter“ Bereiche im Bild Visual Attention for Rapid Scene Analysis Literatur: – – – Itti, Koch, Niebur; A Model of Saliency-Based Visual Attention for Rapid Scene Analysis, IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, VOL. 20, NO. 11, NOVEMBER 1998 Hou, Zhang Department of Computer Science; Saliency Detection: A Spectral Residual Approach, Computer Vision and Pattern Recognition, 2007. CVPR '07. IEEE Conference on Itti, Koch, A saliency-based search mechanism for overt and covert shifts of visual attention, Vision Research 2000 Einzusetzende Mittel: Objekterkennung, Spektralanalyse Robust Energy Based Image Segmentation Methods Robuste Segmentierung bei verrauschten Bildern… … komplexen Strukturen … … und unscharfen Konturen Robust Energy Based Image Segmentation Methods Literatur: – – – Krinidis, Chatzis; Fuzzy Energy-Based Active Contours; IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 18, NO. 12, DECEMBER 2009 Shyu, Pham, Tran, Lee; Global and local fuzzy energy-based active contours for image segmentation Chan, Vese; Active Contours Without Edges, IEEE TRANSACTIONS ON IMAGE PROCESSING, VOL. 10, NO. 2, FEBRUARY 2001 Einzusetzende Mittel: Differentialgleichungen, Segmentierung, Fuzzymethoden Interactive Graph Based Image Segmentation Interaktive Vorgabe von Regionen mit Vordergrund- und Hintergrundtextur Interactive Graph Based Image Segmentation Literatur: – – BOYKOV FUNKA-LEA; Graph Cuts and Efficient N-D Image Segmentation; International Journal of Computer Vision 2006 Li, Fan, Zhao, Zhang; Graph Cuts Based Image Segmentation Using Local Color And Texture; 2011 4th International Congress on Image and Signal Processing Einzusetzende Mittel: Graphalgorithmen, Textursegmentierung IMAGE FUSION High Dynamic Range Imaging High Dynamic Range Imaging Literatur: – – – – – – – – Debevec, Malik; Recovering High Dynamic Range Radiance Maps from Photographs; ACM 1997 Robertson, Borman, Stevenson; Dynamic Range Improvement Through Multiple Exposures; IEEE 1999 Robertson, Borman, Stevenson; Estimation-theoretic approach to dynamic range enhancement using multiple exposures; 2003 Mann; Extending Dynamic Range By Combining Differently Exposed Pictures; IS&T 1994 Healey; Radiometric CCD Camera Calibration and Noise Estimation; IEEE 1994 Mertens, Kautz, Van Reeth; Exposure Fusion; Greg Ward; Fast, robust image registration for compositing high dynamic range photographcs from hand-held exposures; Nayar, Mitsunaga; High dynamic range imaging spatially varying pixel exposures; Einzusetzende Mittel: Least-Squares, Polynomiale Optimierung Fusion of Stereo and Focus Series for Depth Estimation Defocus → Stereo → Detail Stereo Depth Map Fused Depth Map Fusion of Stereo and Focus Series for Depth Estimation Literatur: - Ioana Gheta, Christian Frese, Michael Heizmann; Fusion of Combined Stereo and Focus Series for Depth Estimation; 2006 - Ioana Gheta; Fusion von Stereo- und Fokusserien; 2006 - Christian Frese and Ioana Gheta; Robust Depth Estimation by Fusion of Stereo and Focus Series Acquired with a Camera Array; 2006 - Johannes Burge and Wilson S. Geisler; Optimal defocus estimation in individual natural images; 2011 - Akira Kubota, Kazuya Kodama, and Kiyoharu Aixawa; REGISTRATION AND BLUR ESTIMATION METHODS FOR MULTIPLE DIFFERENTLY FOCUSED IMAGES; 1999 - Gunther Schneider, Bernard Heit, Johannes Honag and Jacques Bremont; MONOCULAR DEPTH PERCEPTION BY EVALUATION OF THE BLUR IN DEFOCUSED IMAGES; 1994 Einzusetzende Mittel: Frequenzanalyse, Differentialrechnung, Optik GRAPHS and MARKOV RANDOM FIELDS FOR OBJECT SEGMENTATION Markov Random Fields (MRFs) in der Bildverarbeitung Degradation Model Anwendungsbeispiel: Textursegmentierung Nachbarschaftsmodell Markov Random Fields (MRFs) in der Bildverarbeitung Literatur: – – – – S. Z. Li; Markov Random Field Modeling in Image Analysis; 2009 Gerhard Winkler; Image Analysis, Random Fields and Markov Chain Monte Carlo Methods: A Mathematical Introduction; 2006 Jun S. Liu; Monte Carlo Strategies in Scientific Computing; 2008 Urs Köster, Jussi T. Lindgren and Aapo Hyvarinen; Estimating Markov Random Field Potentials for Natural images; 2009 Einzusetzende Mittel: Wahrscheinlichkeitstheorie, Mustererkennung State-of-the-Art Object Segmentation Input Boundaries Image Formation Model: Patch-based Classifier Local Consistency Result State-of-the-Art Object Segmentation Literatur: Diane Larlus, Jakob Verbeek and Frédéric Jurie; Category Level Object Segmentation by Combining Bag-of-Words Models with Dirichlet Processes and Random Fields; IEEE Int. J. Computer Vision, 2009 Einzusetzende Mittel: Merkmalsextraktion, Markov Random Fields, Zufallsprozesse, Modellbildung Object Cosegmentation Segmenting the Same Object in Multiple Images Object Cosegmentation Segmenting the Same Object in Multiple Images Literatur: - Carsten Rother, Vladimir Kolmogorov, Tom Minka and Andrew Blake; Cosegmentation of Image Pairs by Histogram Matching — Incorporating a Global Constraint into MRFs; - Sara Vicente, Carsten Rother, Vladimir Kolmogorov; Object Cosegmentation; - Sara Vicente, Vladimir Kolmogorov, Carsten Rother; Cosegmentation Revisited: Models and Optimization; ECCV 2010 - Dorit S. Hochbaum, Vikas Singh; An efficient algorithm for Cosegmentation; Einzusetzende Mittel: Merkmalsextraktion, Markov Random Fields, Graph-Matching Multi-Object Tracking – Walking Pyramids Pyramid of Moving Object Multi-Object Tracking – Walking Pyramids Literatur: Walter Kropatsch: http://www.prip.tuwien.ac.at/people/krw/more/bibliography.php Walter Kropatsch; When Pyramids Learned Walking; Springer 2009 Walter Kropatsch; Representing Scenes with dynamic objects by Graph Pyramids; IASTED 2010 Einzusetzende Mittel: Merkmalsextraktion, Graphen, Geometrische Transformationen MACHINE LEARNING Digital Face Beautification Original Beautified Digital Face Beautification Literatur: Tommer Leyvand, Daniel Cohen-Or, Gideon Dror and Dani Lischinski; Data-Driven Enhancement of Facial Attractiveness; ACM SIGGRAPH 2008 Einzusetzende Mittel: Merkmalsextraktion, Neuronale Netze, Bildregistrierung, Modellbildung Evolutionary Learning of Collaborative Behavior Two teams of robots communicating through light signals for collaboratively finding food but avoiding poison „Robotic Sand“: Tiny robots communicating with their neighbours through magnets for self-assembly of learned 3D shapes Evolutionary Learning of Collaborative Behavior Literatur: - Dario Floreano, Sara Mitri, Stephane Magnenat, Laurent Keller; Evolutionary Conditions for the Emergence of Communication in Robots; 2007 - Richard A. Watson, Sevan G. Ficici, Jordan B. Pollack; Embodied Evolution: Embodying an Evolutionary Algorithm in a Population of Robots; 1999 - Kyle Gilpin and Daniela Rus; Self-disassembling Robot Pebbles: New results and ideas for self-assembly of 3d structures; IEEE 2010 Distributed Robotics Group: http://groups.csail.mit.edu/drl/wiki/index.php?title=Robot_Pebbles Einzusetzende Mittel: Genetische Algorithmen, Knowledge Engineering RGB-D SLAM Simultaneous localization and mapping SLAM with a RGB-D camera voxel grid: SLAM with a RGB-D camera Literatur: - Nikolas Engelhard, Felix Endres, Jürgen Hess, Jürgen Sturm, Wolfram Burgard; Real-time 3D visual SLAM with a hand-held RGB-D camera; 2011 - F. Endres, J. Hess, N. Engelhard, J. Sturm, D. Cremers, W. Burgard; An Evaluation of the RGB-D SLAM System; 2012 - Virgile Högman; Building a 3D map from RGB-D sensors; 2012 - Software quelloffen (http://www.ros.org/wiki/rgbdslam) Einzusetzende Mittel: lokale Bildmerkmale, (Probabilistik, Optimierung) Generalized-ICP (Iterative Closest Point) Generalized-ICP (Iterative Closest Point) Literatur: - Aleksandr V. Segal, Dirk Haehnel, Sebastian Thrun; Generalized-ICP; 2009 - Weitere Verweise im Paper (ICP und point-to-plane ICP) - Implementiert in der Point Cloud Library (PCL) Einzusetzende Mittel: Optimierung, Probabilistik, geometrische Transformationen Hierarchical Optimization for Online Mapping Hierarchical Optimization for Online Mapping Literatur: - Giorgio Grisetti, Rainer Kümmerle, Cyrill Stachniss, Udo Frese, Christoph Hertzberg; Hierarchical Optimization on Manifolds for Online 2D and 3D Mapping; 2010 - Weitere Verweise im Paper - quelloffen („HOG-Man“) Einzusetzende Mittel: Optimierung (Gauß-Newton-Verfahren), Geometrische Transformationen (lineare Näherung bezüglich Rotation) SLAM with „general graph optimization“ SLAM with „general graph optimization“ Literatur: - Rainer Kümmerle, Giorgio Grisetti, Hauke Strasdat, Kurt Konolige, Wolfram Burgard; g2o: A General Framework for Graph Optimization; 2011 - Weitere Verweise im Paper - quelloffen („g2o“) Einzusetzende Mittel: Optimierung (diverses), Geometrische Transformationen (lineare Näherung bezüglich Rotation) Themenliste 1) Automatic Panorama Image Stitching 2) Visual Attention for Rapid Scene Analysis 3) Robust Energy Based Image Segmentation Methods 4) Interactive Graph Based Image Segmentation 5) High Dynamic Range Imaging 6) Fusion of Stereo and Focus Series for Depth Estimation 7) Markov Random Fields (MRFs) in der Bildverarbeitung 8) State-of-the-Art Object Segmentation 9) Object Cosegmentation 10) Multi-Object Tracking – Walking Pyramids 11) Digital Face Beautification 12) Evolutionary Learning of Collaborative Behavior 13) SLAM with a RGB-D camera 14) Generalized-ICP 15) Hierarchical Optimization for Online Mapping 16) SLAM with „general graph optimization“