Digitales Sehen mit Bildpyramiden (bibtex)
by Walter G. Kropatsch
Abstract:
Digital Computers are nowadays capable to acquire and to process visual information. The research area of 'Computer Vision' has as its goal to teach the machine abilities comparable to the human eye. At first we shall consider characteristical performance achieved by natural vision in order to derive constraints for digital vision.Then the analysis of digital images is classified into different classes of processing steps ordered according to the different degree of abstraction of the involved descriptions. An image pyramid is presented as a hierarchical data structure which is able to contain both images of various resolutions as well as image descriptions of various degrees of abstraction. Several realizations of pyramidal systems do exist and have shown that simple tasks of image analysis can be performed in real time. In dual pyramids a curve pyramid complements a pyramid of images. Some examples of this concept indicate the potential that the model of the pyramid still offers for computer vision.
Reference:
Digitales Sehen mit Bildpyramiden (Walter G. Kropatsch), Technical report, PRIP, TU Wien, 1991.
Bibtex Entry:
@TechReport{TR001,
  author =       "Walter G. Kropatsch",
  institution =  "PRIP, TU Wien",
  number =       "PRIP-TR-001",
  title =        "Digitales {S}ehen mit {B}ildpyramiden",
  year =         "1991",
  price =        "35",
  url =          "ftp://ftp.prip.tuwien.ac.at/pub/publications/trs/tr1.ps.gz",
  abstract =     "Digital Computers are nowadays capable to acquire and
                 to process visual information. The research area of
                 'Computer Vision' has as its goal to teach the machine
                 abilities comparable to the human eye. At first we
                 shall consider characteristical performance achieved by
                 natural vision in order to derive constraints for
                 digital vision.Then the analysis of digital images is
                 classified into different classes of processing steps
                 ordered according to the different degree of
                 abstraction of the involved descriptions. An image
                 pyramid is presented as a hierarchical data structure
                 which is able to contain both images of various
                 resolutions as well as image descriptions of various
                 degrees of abstraction. Several realizations of
                 pyramidal systems do exist and have shown that simple
                 tasks of image analysis can be performed in real time.
                 In dual pyramids a curve pyramid complements a pyramid
                 of images. Some examples of this concept indicate the
                 potential that the model of the pyramid still offers
                 for computer vision.",
}
Powered by bibtexbrowser