• DocumentCode
    665117
  • Title

    Catadioptric omnidirectional image inpainting via a multi-scale approach and image unwrapping

  • Author

    Paredes, Daniel ; Rodriguez, Paul ; Ragot, N.

  • Author_Institution
    Electr. Dept., Pontificia Univ. Catolica del Peru, Lima, Peru
  • fYear
    2013
  • fDate
    21-23 Oct. 2013
  • Firstpage
    67
  • Lastpage
    72
  • Abstract
    Omnidirectional catadioptric sensors are widely used in different applications as they allow the observation of a 360 field of view instantaneously. Their specific architecture necessarily induces carrying devices which obstruct the field of view of the sensor. This drawback can be a major obstacle to their use. This paper outlines a methodoly to inpaint catadioptric omnidirectional images, that means to remove and repair a part of the image. This method combines a multi-scale image inpainting (MII) algorithm with unwrapping techniques as the MII can not be applied in the omnidirectional image directly.
  • Keywords
    Markov processes; image resolution; image sensors; MII algorithm; catadioptric omnidirectional image inpainting; image repair; image unwrapping; multiscale approach; multiscale image inpainting algorithm; omnidirectional catadioptric sensors; unwrapping techniques; Biosensors; Cameras; Image resolution; Markov random fields; Mirrors; Visualization;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Robotic and Sensors Environments (ROSE), 2013 IEEE International Symposium on
  • Conference_Location
    Washington, DC
  • Print_ISBN
    978-1-4673-2938-5
  • Type

    conf

  • DOI
    10.1109/ROSE.2013.6698420
  • Filename
    6698420