• DocumentCode
    678847
  • Title

    Estimation of Two Illuminant Spectral Power Distributions from Highlights of Overlapping Illuminants

  • Author

    Nguyen Thuc Dieu Hang ; Horiuchi, T.K. ; Hirai, Keita ; Tominaga, Shoji

  • Author_Institution
    Grad. Sch. of Adv. Integration Sci., Chiba Univ., Chiba, Japan
  • fYear
    2013
  • fDate
    2-5 Dec. 2013
  • Firstpage
    434
  • Lastpage
    440
  • Abstract
    A method is proposed for estimating two illuminant spectral power distributions from the highlights of overlapping illuminants on an object. It was assumed in most previous studies that specular highlights on object surfaces are caused by a single light source, or by separate multiple light sources. In this paper, we assume that two highlights from different light sources are overlapped on object surfaces, and estimate the spectral power distributions of both. A multiband camera system is used for capturing spectral images of dielectric objects in a scene. First, we detect specular highlight areas from the spectral image. Then, the illuminant spectra of two light sources are simultaneously estimated based on cluster classification of particular pixel distribution in the highlight area. The feasibility of the proposed method is examined in experiments on real-world scenes.
  • Keywords
    computer vision; image classification; image sensors; lighting; cluster classification; dielectric objects; illuminant spectral power distributions; multiband camera system; multiple light sources; object surfaces; overlapping illuminants; pixel distribution; spectral image; specular highlight areas; specular highlights; Cameras; Estimation; Image color analysis; Light sources; Power distribution; Reflection; Vectors; illuminant estimation; mean-shift algorithm; multiple light sources; specular highlight area;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Signal-Image Technology & Internet-Based Systems (SITIS), 2013 International Conference on
  • Conference_Location
    Kyoto
  • Type

    conf

  • DOI
    10.1109/SITIS.2013.77
  • Filename
    6727226