• DocumentCode
    2101520
  • Title

    The Ghosting Technique Applied to Augmented Reality Visualization

  • Author

    Padilha, Arthur ; Rolim, Catarina ; Teichrieb, Veronica

  • Author_Institution
    Voxar Labs. Centro de Inf. - UFPE, Recife, Brazil
  • fYear
    2013
  • fDate
    28-31 May 2013
  • Firstpage
    159
  • Lastpage
    166
  • Abstract
    Functional realism focuses on delivering to the user the best visual information from the target scene. The virtual scenes/objects do not need to behave or look like the real ones, but they need to help users in tasks execution. In this sense, this paper focuses on functional realism applied to Augmented Reality scenes. The technique presented here applies a visualization technique called Ghosting to generate an importance map from the real scene in order to blend in the virtual objects in a more adequate way. A careless augmentation can prejudice the user Augmented Reality based experience. The importance map is built from a saliency map, edge detection, super pixel representation and texture analysis. We present results in outdoor and indoor environments in order to prove the technique effectiveness.
  • Keywords
    augmented reality; data visualisation; edge detection; image representation; image texture; natural scenes; augmented reality scene visualization; edge detection; functional realism; ghosting technique; importance map generation; indoor environments; outdoor environments; real scene; saliency map; superpixel representation; target scene; texture analysis; user augmented reality-based experience; virtual objects; virtual scenes; visual information; Abstracts; Art; Augmented reality; Image edge detection; Image segmentation; Pipelines; Visualization; Augmented Reality; Ghosting Technique; Saliency Map; Transparency; Visualization;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Virtual and Augmented Reality (SVR), 2013 XV Symposium on
  • Conference_Location
    Cuiaba
  • Type

    conf

  • DOI
    10.1109/SVR.2013.36
  • Filename
    6655774