Extraction of Glasses in Human Face Images

  • Yi Xiao
  • Hong Yan
Part of the Lecture Notes in Computer Science book series (LNCS, volume 3072)


In this paper, a face model is established based on the Delaunay triangulation for extraction of eyeglasses in human face images. A face shape is decomposed into a few chains represented by the non-shortest edge shared triangles. The width and length of a chain are quantitatively measured by the triangles in the chain. An eye region consists of small sized triangles and triangle chains with short length and narrow width. And an eye region and its glasses frame are linked the triangle chains called surrounding area. Eyes and the glasses are located by measuring the symmetry and similarity of any two regions and their surrounding areas. Experiments show that the glass boundary in various shapes can be traced automatically and accurately and small broken edges can be linked without obviously distortion.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Jing, Z., Mariani, R., Wu, J.: Glasses detection and extraction by deformable contour. In: Proceedings of the 15th International Conference on Pattern Recognition, vol. II, p. 4B (2000)Google Scholar
  2. 2.
    Jing, Z., Mariani, R., Wu, J.: Glasses Detection and Extraction by Bayes Rules and Deformable Contour. In: The 4th world multiconference on Systemics, Cybernetics and Informatics (SCI), Orlando, Florida (2000)Google Scholar
  3. 3.
    Wu, C., Liu, C., Shum, H.Y., Xu, Y.Q., Zhang, Z.: Automatic Eyeglasses Removal from Face Images. In: ACCV 2002: The 5th Asian Conference on Computer Vision, Australia (2002)Google Scholar
  4. 4.
    Wu, H., Yoshikawa, G., Shioyama, T., Lao, T., Kawade, T.: Glasses frame detection with 3D Hough transform. In: Proceedings. 16th International Conference on Pattern Recognition, vol. 2, pp. 346–349 (2002)Google Scholar
  5. 5.
    Okabe, A., Boots, B., Sugihara, K.: Spatial Tessellations-Concepts and Applications of Voronoi Diagrams, 2nd edn. Wiley Series in probability and statistics (2000)Google Scholar
  6. 6.
    Xiao, Y., Yan, H.: Text region extraction in a document image based on the Delaunay tessellation. Pattern Recognition 36(3), 799–809 (2003)zbMATHCrossRefMathSciNetGoogle Scholar
  7. 7.

Copyright information

© Springer-Verlag Berlin Heidelberg 2004

Authors and Affiliations

  • Yi Xiao
    • 1
  • Hong Yan
    • 1
    • 2
  1. 1.School of Electrical & Information EngineeringUniversity of SydneyAustralia
  2. 2.Department of Computer Engineering & Information TechnologyCity University of Hong KongKowlon, Hong Kong

Personalised recommendations