Jaklič A., Leonardis A., Solina F. Segmentation and Recovery of Superquadrics

Издательство Kluwer, 2000, -273 pp.In recent years, superquadrics established themselves as a popular model for representation of objects and 3D scenes in computer vision, computer graphics, and robotics. Superquadrics are a family of parametric models that cover a wide variety of smoothly changing 3D shapes which are controlled with a small number of parameters. For greater versatility the superquadric shapes can be augmented with global and local deformations. Superellipsoids, a subset of superquadrics, are volumetric models particularly suited for part-level modeling of 3D scenes which directly supports reasoning and manipulation. Especially, interpretation of range images in computer vision has been infiuenced by various methods for recovery of superquadric models from image data.
Readers of this book can expect to find a thorough evolution and definition of superquadric models, as well as derivations of their various geometric properties. Advantages and disadvantages of superquadrics in comparison to other volumetric models used in computer vision are addressed. Applications of superquadrics in computer vision and robotics are thoroughly discussed and in particular, the use of superquadrics for range image registration is demonstrated.
The central theme of the whole book is our method of recovery and segment at ion of superquadrics from range images. The method is in essence the result of doctoral dissertations of all three authors which span a nearly ten year period. It is described in detail and compared with other methods of superquadric recovery and segmentation. Numerous examples of recovery and segmentation from range images are given.
The intended audience for this book are researchers, developers, and students of computer vision and robotics. It is assumed that the readers have a general engineering/computer science background with some familiarity of computer vision issues.Superquadrics and Their Geometric Properties
Extensions of Superquadrics
Recovery of Individual Superquadrics
Segmentation with Superquadrics
Experimental Results
Applications of Superquadrics
Conclusions
A: Rendering of Superquadrics in Mathematica
B: Superquadric Recovery Code
C: Range Image Acquisition
D: Minimum Description Length and Maximum A Posteriori Probability
E: Object-Oriented Framework for Segmentation (Segmentor)