• DocumentCode
    3081528
  • Title

    A Perceptual Metric Based on Salient Information Entropy for 3D Mesh Distortion

  • Author

    Shi, Zhenfeng ; Li, Qiong ; Niu, Xiamu

  • Author_Institution
    Sch. of Comput. Sci. & Technol., Harbin Inst. of Technol., Harbin, China
  • fYear
    2010
  • fDate
    15-17 Oct. 2010
  • Firstpage
    563
  • Lastpage
    566
  • Abstract
    Visual degradation and distortion may be introduced during the 3D mesh simplification. The critical issue in all mesh simplification methods is how to evaluate it. Some geometric metrics without the consideration of human visual perception and perceptual metrics based on subjective test have been proposed. It is urgent to develop an objective evaluation metric with the consideration of human visual perception. We used mesh saliency integrating with Just-Noticeable Difference(JND) methodology to define the salient information entropy based on the theory of information and presented a perceptual metric based on it for measuring the visual degradation and distortion introduced by 3D mesh simplification. Experimental results show that our perceptual metric can evaluate the visual degradation effectively and the essential of 3D mesh simplification is decreasing the visual redundancy and average visual uncertainty.
  • Keywords
    computational geometry; entropy; mesh generation; solid modelling; visual perception; 3D mesh distortion; geometric metrics; human visual perception; just noticeable difference; perceptual metrics; salient information entropy; visual degradation; visual redundancy; Degradation; Humans; Information entropy; Measurement; Three dimensional displays; Visual perception; Visualization; HVS; distortion; entropy; perception; saliency;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Intelligent Information Hiding and Multimedia Signal Processing (IIH-MSP), 2010 Sixth International Conference on
  • Conference_Location
    Darmstadt
  • Print_ISBN
    978-1-4244-8378-5
  • Electronic_ISBN
    978-0-7695-4222-5
  • Type

    conf

  • DOI
    10.1109/IIHMSP.2010.143
  • Filename
    5635561