Reconnaissance d'objets utilisant des histogrammes multidimensionnels de champs réceptifs. (Object Recognition using Multidimensional Receptive Field Histograms)

Au cours des dernieres annees, l'interet pour des algorithmes de reconnaissance fondes sur l'apparence a considerablement augmente. Ces algorithmes utilisent directement des informations d'images. A partir des images d'objets ces approches construisent des "modeles fondes sur l'apparence", car chaque image representee correspond a une apparence particuliere d'un objet. La fiabilite, la vitesse et le taux de reconnaissance eleve de ces techniques en constituent les interets majeurs. Le succes de ces methodes est considerable pour la reconnaissance de visages, dans le contexte de l'interface homme-machine et pour l'acces a des bases d'images par leurs contenus. Cette these propose une technique ou les objets sont representes par des statistiques sur des operateurs locaux et robustes. On veut montrer qu'une telle representation fondee sur l'apparence est fiable et extremement discriminante pour la reconnaissance d'objets. La motivation initiale de cette etude etait la reconnaissance rapide d'objets par la methode des histogrammes de couleurs. Cette methode utilise les statistiques de couleurs comme modele d'objets. La premiere partie de la these generalise cette approche en modelisant des objets par les statistiques de leurs caracteristiques locales. La technique generalisee - que l'on appelle "histogrammes multidimensionnels de champs" receptifs - permet de discriminer un grand nombre d'objets. Les faiblesses de cette approche sont liees aux "defis des modeles fondes sur l'apparence". Ces defis concernent la reconnaissance en presence d'occultation partielle, la reconnaissance d'objets 3D a partir des images 2D et la classification d'objets comme generalisation en dehors de la base d'objets. La deuxieme partie de la these examine chacun de ces defis et propose une extension appropriee de notre technique. L'interet principal de cette these est le developpement d'un modele de representation d'objets qui utilise les statistiques de vecteurs de champs receptifs. Plusieurs algorithmes de l'identification et aussi de la classification d'objets sont proposes. En particulier, un algorithme probabiliste est defini : il ne depend pas de la correspondance entre les images de test et les objets de la base de donnees. Des experiences obtiennent des taux de reconnaissance eleves en utilisant le modele de representation propose. dans un etat de l'art on decrit brievement des techniques qui ont etes sources d'inspiration : des techniques de histogrammes de couleurs, des algorithmes de reconnaissance fondee sur des descripteurs locaux et des approches de la representation et reconnaissance statistique d'objet. Afin de generaliser la technique de la comparaison d'histogrammes de couleurs, des descripteurs locaux sont discutes. Differentes techniques de comparaison d'histogrammes sont proposees et leur robustesse par rapport au bruit et au changement de l'intensite d'eclairage est analysee. Dans des experimentations de l'identification d'une centaine d'objets les differents degres de liberte de la reconnaissance d'objets sont consideres : changements d'echelle et de la rotation d'image, variations du point de vue et occultation partielle. Un algorithme probabiliste est propose, qui ne depend pas de la correspondance entre les images de test et les objets de la base de donnees. Des experiences obtiennent des taux de reconnaissance eleves en utilisant seulement une petite partie visible d'objet. Enfin une extension de cet algorithme fondee sur une table de hachage dynamique est proposee pour la reconnaissance de plusieurs objets dans les scenes complexes. Deux algorithmes actifs de reconnaissance d'objets sont proposes. Un algorithme calcule des regions d'interet pour le controle de fixation d'une camera en 2D. Le deuxieme algorithme propose la planification de points de vue pour la reconnaissance des objets 3D a partir des apparence d'objets en 2D. Un dernier chapitre propose le concept des classes visuelles definis par des similarites d'objets comme cadre general pour la classification d'objets. Une technique selon le maximum de vraisemblance est propose pour la reconnaissance des classes visuelles et appliquee pour obtenir des images visuellement similaire d'une base d'images.

[1]  G. Healey,et al.  Using Illumination Invariant Color Histogram Descriptors for Recognit ion , 1994 .

[2]  Graham D. Finlayson,et al.  Color constancy in diagonal chromaticity space , 1995, Proceedings of IEEE International Conference on Computer Vision.

[3]  Rakesh Mohan,et al.  Systematic design of indexing strategies for object recognition , 1993, Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.

[4]  Azriel Rosenfeld,et al.  Image analysis: Problems, progress and prospects , 1984, Pattern Recognit..

[5]  Shimon Edelman,et al.  On Learning to Recognize 3-D Objects from Examples , 1993, IEEE Trans. Pattern Anal. Mach. Intell..

[6]  John F. Canny,et al.  A Computational Approach to Edge Detection , 1986, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[7]  Richard A. Young,et al.  SIMULATION OF HUMAN RETINAL FUNCTION WITH THE GAUSSIAN DERIVATIVE MODEL. , 1986 .

[8]  Bernt Schiele,et al.  Where to look next and what to look for , 1996, Proceedings of IEEE/RSJ International Conference on Intelligent Robots and Systems. IROS '96.

[9]  Daniel P. Huttenlocher,et al.  Computing visual correspondence: incorporating the probability of a false match , 1995, Proceedings of IEEE International Conference on Computer Vision.

[10]  Gérard G. Medioni,et al.  Finding Waldo, or focus of attention using local color information , 1993, Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.

[11]  C. NelsonDepartment,et al.  3-d Recognition via 2-stage Associative Memory , 1995 .

[12]  L Sirovich,et al.  Low-dimensional procedure for the characterization of human faces. , 1987, Journal of the Optical Society of America. A, Optics and image science.

[13]  Takeo Kanade,et al.  Region segmentation: Signal vs semantics , 1980 .

[14]  Ruzena Bajcsy,et al.  Combining color and geometry for the active, visual recognition of shadows , 1995, Proceedings of IEEE International Conference on Computer Vision.

[15]  Alex Waibel,et al.  Gaze Tracking Based on Face‐Color , 1995 .

[16]  David G. Lowe,et al.  Learning to recognize objects in images: acquiring and using probabilistic models of appearance , 1995 .

[17]  Patrick Gros,et al.  Rapid Object Indexing and Recognition Using Enhanced Geometric Hashing , 1996, ECCV.

[18]  M. Turk,et al.  Eigenfaces for Recognition , 1991, Journal of Cognitive Neuroscience.

[19]  H. Nagel Reflections on Active (Machine) Vision , 1992 .

[20]  Bernt Schiele,et al.  Transinformation of object recognition and its application to viewpoint planning , 1997, Robotics Auton. Syst..

[21]  L. Maloney,et al.  Color constancy: a method for recovering surface spectral reflectance. , 1986, Journal of the Optical Society of America. A, Optics and image science.

[22]  Max A. Viergever,et al.  General Intensity Transformations and Second Order Invariants , 1992 .

[23]  W. Eric L. Grimson,et al.  On the Sensitivity of the Hough Transform for Object Recognition , 1990, IEEE Trans. Pattern Anal. Mach. Intell..

[24]  Jérôme Martin,et al.  An Appearance-Based Approach to Gesture-Recognition , 1997, ICIAP.

[25]  Michael J. Swain,et al.  Interactive indexing into image databases , 1993, Electronic Imaging.

[26]  W. Eric L. Grimson,et al.  Introduction to the Special Issue on Interpretation of 3-D Scenes-Part I , 1991, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[27]  Kenji Nagao,et al.  Recognizing 3D objects using photometric invariant , 1995, Proceedings of IEEE International Conference on Computer Vision.

[28]  Masud Mansuripur,et al.  Introduction to information theory , 1986 .

[29]  Michael J. Swain,et al.  The capacity of color histogram indexing , 1994, 1994 Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.

[30]  James Lee Hafner,et al.  Efficient Color Histogram Indexing for Quadratic Form Distance Functions , 1995, IEEE Trans. Pattern Anal. Mach. Intell..

[31]  Azriel Rosenfeld,et al.  Computer Vision , 1988, Adv. Comput..

[32]  J. Hornegger,et al.  Statistical learning, localization, and identification of objects , 1995, Proceedings of IEEE International Conference on Computer Vision.

[33]  H. Martin Hunke,et al.  Locating and Tracking of Human Faces with Neural Networks , 1994 .

[34]  Randal C. Nelson Memory-Based Recognition for 3-D Objects , 1996 .

[35]  Brian V. Funt,et al.  Color Constant Color Indexing , 1995, IEEE Trans. Pattern Anal. Mach. Intell..

[36]  David G. Lowe,et al.  Learning indexing functions for 3-D model-based object recognition , 1994, 1994 Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.

[37]  Yehezkel Lamdan,et al.  On recognition of 3-D objects from 2-D images , 2011, Proceedings. 1988 IEEE International Conference on Robotics and Automation.

[38]  Katsushi Ikeuchi,et al.  Why aspect graphs are not (yet) practical for computer vision , 1991, [1991 Proceedings] Workshop on Directions in Automated CAD-Based Vision.

[39]  Jitendra Malik,et al.  A computational model of texture segmentation , 1989, Proceedings CVPR '89: IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[40]  Jiri Matas,et al.  On representation and matching of multi-coloured objects , 1995, Proceedings of IEEE International Conference on Computer Vision.

[41]  Bernt Schiele,et al.  Probabilistic object recognition using multidimensional receptive field histograms , 1996, Proceedings of 13th International Conference on Pattern Recognition.

[42]  Rajesh P. N. Rao,et al.  Seeing Behind Occlusions , 1994, ECCV.

[43]  Kris Popat,et al.  Cluster-based probability model applied to image restoration and compression , 1994, Proceedings of ICASSP '94. IEEE International Conference on Acoustics, Speech and Signal Processing.

[44]  Robert A. Hummel,et al.  Distributed Bayesian object recognition , 1993, Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.

[45]  Brian V. Funt,et al.  Color Angular Indexing , 1996, ECCV.

[46]  Glenn Healey,et al.  Combining color and geometric information for the illumination invariant recognition of 3-D objects , 1995, Proceedings of IEEE International Conference on Computer Vision.

[47]  David G. Lowe,et al.  Learning Appearance Models for Object Recognition , 1996, Object Representation in Computer Vision.

[48]  Rajesh P. N. Rao,et al.  Object indexing using an iconic sparse distributed memory , 1995, Proceedings of IEEE International Conference on Computer Vision.

[49]  Carl F. R. Weiman,et al.  Efficient discrete Gabor functions for robot vision , 1994, Defense, Security, and Sensing.

[50]  Emanuele Trucco,et al.  Geometric Invariance in Computer Vision , 1995 .

[51]  John Daugman,et al.  High Confidence Visual Recognition of Persons by a Test of Statistical Independence , 1993, IEEE Trans. Pattern Anal. Mach. Intell..

[52]  Leonidas J. Guibas,et al.  Image Retrieval and Robot Vision Research at Stanford , 1996 .

[53]  Zane W. Bell A Bayesian/Monte Carlo Segmentation Method for Images Dominated by Gaussian Noise , 1989, IEEE Trans. Pattern Anal. Mach. Intell..

[54]  Jitendra Malik,et al.  A Computational Framework for Determining Stereo Correspondence from a Set of Linear Spatial Filters , 1991, ECCV.

[55]  David G. Lowe,et al.  Learning object recognition models from images , 1993, 1993 (4th) International Conference on Computer Vision.

[56]  Dmitry B. Goldgof,et al.  The Scale Space Aspect Graph , 1992, IEEE Trans. Pattern Anal. Mach. Intell..

[57]  Yehezkel Lamdan,et al.  Geometric Hashing: A General And Efficient Model-based Recognition Scheme , 1988, [1988 Proceedings] Second International Conference on Computer Vision.

[58]  R. Deriche Recursively Implementing the Gaussian and its Derivatives , 1993 .

[59]  Luc Florack Local Image Structure , 1997 .

[60]  Hans Knutsson,et al.  Preattentive gaze control for robot vision , 1992 .

[61]  Markus A. Stricker,et al.  Color and Geometry as Cues for Indexing , 1992 .

[62]  J.B. Burns,et al.  View Variation of Point-Set and Line-Segment Features , 1993, IEEE Trans. Pattern Anal. Mach. Intell..

[63]  Rosalind W. Picard A Society of Models for Video and Image Libraries , 1996, IBM Syst. J..

[64]  Alex Waibel,et al.  Estimation of the Head Orientation Based on a Face‐Color‐Intensifier , 1995 .

[65]  Cordelia Schmid,et al.  Bayesian Decision Versus Voting for Image Retrieval , 1997, CAIP.

[66]  Katsushi Ikeuchi,et al.  Recognition of the multi specularity objects for bin-picking task , 1996, Proceedings of IEEE/RSJ International Conference on Intelligent Robots and Systems. IROS '96.

[67]  Dragutin Petkovic,et al.  Query by Image and Video Content: The QBIC System , 1995, Computer.

[68]  Thomas M. Breuel,et al.  Higher-order statistics in object recognition , 1993, Proceedings of IEEE Conference on Computer Vision and Pattern Recognition.

[69]  Lawrence Sirovich,et al.  Application of the Karhunen-Loeve Procedure for the Characterization of Human Faces , 1990, IEEE Trans. Pattern Anal. Mach. Intell..

[70]  F. E. Vasiliuk The Structure of an Image , 1995 .

[71]  Bartlett W. Mel SEEMORE: Combining Color, Shape, and Texture Histogramming in a Neurally Inspired Approach to Visual Object Recognition , 1997, Neural Computation.

[72]  Michael Georgiopoulos,et al.  Learning geometric hashing functions for model-based object recognition , 1995, Proceedings of IEEE International Conference on Computer Vision.

[73]  Brian V. Funt,et al.  Color constancy under varying illumination , 1995, Proceedings of IEEE International Conference on Computer Vision.

[74]  S. Nayar,et al.  Learning and Recognition of 3-D Objects from Brightness Images * , 1993 .

[75]  Bernt Schiele,et al.  Transinformation for active object recognition , 1998, Sixth International Conference on Computer Vision (IEEE Cat. No.98CH36271).

[76]  Haim Schweitzer,et al.  Learnable and Nonlearnable Visual Concepts , 1990, IEEE Trans. Pattern Anal. Mach. Intell..

[77]  Nathan Intrator,et al.  Three-Dimensional Object Recognition Using an Unsupervised BCM Network: The Usefulness of Distinguishing Features , 1993, Neural Computation.

[78]  Rajesh P. N. Rao,et al.  An Active Vision Architecture Based on Iconic Representations , 1995, Artif. Intell..

[79]  Cordelia Schmid,et al.  Combining greyvalue invariants with local constraints for object recognition , 1996, Proceedings CVPR IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[80]  Cordelia Schmid,et al.  Appariement d'images par invariants locaux de niveaux de gris. Application à l'indexation d'une base d'objets. (Image matching by local greyvalue invariants. Applied to indexing an object database) , 1996 .

[81]  J. Koenderink,et al.  Invariant Third Order Properties of Isophotes: T-Junction Detection , 1992 .

[82]  Bernt Schiele,et al.  The Concept of Visual Classes for Object Classification , 1997 .