Representation and recovery of 3D curved objects using generalized cylinders and the extended gaussian image

Abstract We present analytic and experimental results of an initial investigation into the generic representation of curved 3D objects from a 2D input projection. The method uses the cross-section, spine and sweeping rule of generalized cylinders, and is obtained using spherical trigonometry and the extended gaussian image. The method combines the advantages of global and local approaches, and provides a means for curved 3D shape recognition and reconstruction. Results are presented using real world images of mug, bowl and spoon, alongside an analysis of stability considerations.

[1]  Ramakant Nevatia,et al.  Recognizing 3-D Objects Using Surface Descriptions , 1989, IEEE Trans. Pattern Anal. Mach. Intell..

[2]  Gérard G. Medioni,et al.  Inference of segmented, volumetric shape from three intensity images , 1996, Proceedings CVPR IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[3]  Ian Naismith Sneddon,et al.  Encyclopaedic dictionary of mathematics for engineers and applied scientists , 1976 .

[4]  Thomas O. Binford,et al.  Inferring Surfaces from Images , 1981, Artif. Intell..

[5]  Azriel Rosenfeld,et al.  Human and Machine Vision , 1983 .

[6]  M. Brady.  Criteria for Representations of Shape , 1983 .

[7]  Xiang Pan,et al.  A parallel method for locating and representing 2D contours , 1992, Pattern Recognit. Lett..

[8]  Xiang Pan.  Approaches for edge detection, pose determination and object representation in computer vision , 1994 .

[9]  Y. J. Tejwani,et al.  Robot vision , 1989, IEEE International Symposium on Circuits and Systems,.

[10]  Yunde Jia.  Description and recognition of curved objects , 1992, Proceedings., 11th IAPR International Conference on Pattern Recognition. Vol. III. Conference C: Image, Speech and Signal Analysis,.

[11]  Berthold K. P. Horn.  Extended Gaussian images , 1984, Proceedings of the IEEE.