A method of red-eye effect recognition in digital photography

Authors

  • A.V. Agayan
  • G.M. Agayan
  • S.Yu. Romanov

Keywords:

распознавание образов
цифровые фотографии
эффект красного глаза
конечнопараметрическая модель

Abstract

The paper is devoted to the red-eye effect recognition on digital photos made with flash. RGB digital photos were used as input data. The recognition algorithm consists of three steps: preliminary, finite parametric filtration, and context filtration. After that the recognition correction is performed. The algorithm was tested on a wide range of different-quality photos. The best recognition was achieved on the photos with large areas of pupils.

New computational technologies

Downloads

Published

2005-10-20

Issue

Vol. 6 (2005): Issue 1.

Section

Section 1. Numerical methods and applications

Author Biographies

A.V. Agayan

Lomonosov Moscow State University,
Research Computing Center

G.M. Agayan

Lomonosov Moscow State University,
Research Computing Center

S.Yu. Romanov

Lomonosov Moscow State University,
Research Computing Center

References

  1. Schildkraut J.S., Gray R.T. A fully automatic red-eye detection and correction algorithm // The 9th International Conference on Image Processing. September 23,-,26. T. I. Rochester, 2002. 801-803.
  2. Верлань А.Ф., Сизиков В.С. Интегральные уравнения: методы, алгоритмы, программы. Киев: Наукова Думка, 1986.
  3. Gaubatz M., Ulichney R. Automatic red-eye detection and correction // The 9th International Conference on Image Processing. September 23,-,26. T. II. Rochester, 2002. 103-107.
  4. Васильев Ф.П. Численные методы решения экстремальных задач. М.: Наука,