Multi-pass Feedback Control for Object Recognition

In order to locate interesting areas of an image we describe a system for focus of attention; this is based on feedback strategies combining low-level features, and a high-level object model to recognise the object and to direct the search for missing information. We aim to improve on established single-pass hypothesis generation and veri cation approaches by applying our complex feedback strategies to recognise generic classes of objects. By using a complex feedback strategy we produce optimal sets of low level features and reduce the number of hypotheses generated. The system can extract simple and complex objects in a scale and rotation independent manner where the objects may be partially occluded. The method is illustrated for simple cubic objects and the results are expected to be applied for a mobile robot application. A n de situer les r egions d'int erêt d'une image, nous d ecrivons un syst eme de centre d'attention bas e sur des strat egies de retour de donn ees. Cellesci comprennent plusieurs caract eristiques de basniveau ainsi q'un mod ele de haut-niveau lequel reconnaît l'objet et dirige l'op eration de recherche des donn ees manquantes. Notre objectif est d'am eliorer l'approche de g en eration et de v eri cation des hypoth eses en utilisant nos strat egies de retour de donn ees a n d'identi er des cat egories g en eriques d'objets. Grâce a ces strat egies nous obtiendrons un ensemble optimal de caract eristiques de bas-niveau et nous r eduirons le nombre d'hypoth eses g en er ees. Le syst eme est capable de reconnaître des objets simples ou complexes ind ependamment de l' echelle, de la rotation ou bien même de l'occlusion partielle d'un objet. Nous d emontrerons cette m ethode pour de simples objets cubiques. Celle-l a sera utilis ee Now at Emrad Ltd., Catteshall Lane, Godalming, Surrey GU7 1NG, UK pour une application de robot mobile.

[1]  W. Eric L. Grimson,et al.  The Combinatorics Of Object Recognition In Cluttered Environments Using Constrained Search , 1988, [1988 Proceedings] Second International Conference on Computer Vision.

[2]  Josef Kittler,et al.  An interest operator based on perceptual grouping , 1996 .

[3]  W. Eric L. Grimson,et al.  On the Verification of Hypothesized Matches in Model-Based Recognition , 1991, IEEE Trans. Pattern Anal. Mach. Intell..

[4]  Shimon Ullman,et al.  Structural Saliency: The Detection Of Globally Salient Structures using A Locally Connected Network , 1988, [1988 Proceedings] Second International Conference on Computer Vision.

[5]  Kok Cheong Wong Representation feature extraction and geometric constraints for recognising 3D objects from a single perspective view. , 1992 .

[6]  Bruce A. Draper,et al.  An Example of Learning in Knowledge-Directed Vision , 1992 .

[7]  Virgil W. Eveleigh,et al.  Adaptive Control And Optimization Techniques , 1967 .

[8]  Tony Lindeberg Detecting salient blob-like image structures and their scales , 1994 .

[9]  Josef Kittler,et al.  Optimal Edge Detectors for Ramp Edges , 1991, IEEE Trans. Pattern Anal. Mach. Intell..

[10]  Charles R. Dyer,et al.  Model-based recognition in robot vision , 1986, CSUR.

[11]  Josef Kittler,et al.  A Hough transform algorithm with a 2D hypothesis testing kernel , 1993 .

[12]  J Kittler,et al.  Hypothesis Generation and Veriication Using Complex Feedback Strategies for Object Recognition Hypothesis Generation and Veriication Using Complex Feedback Strategies for Object Recognition , 2007 .

[13]  Josef Kittler,et al.  A Performance Measure for Boundary Detection Algorithms , 1996, Comput. Vis. Image Underst..

[14]  David G. Lowe,et al.  Perceptual Organization and Visual Recognition , 2012 .

[15]  J. K. Aggarwal,et al.  Detection and segmentation of man-made objects in outdoor scenes: concrete bridges , 1989 .

[16]  Jean Ponce,et al.  Probabilistic 3D object recognition , 1995, Proceedings of IEEE International Conference on Computer Vision.

[17]  Jiri Matas,et al.  Junction detection using probabilistic relaxation , 1993, Image Vis. Comput..