Security

Colour Invariants for Machine Face Recognition

Free registration required

Executive Summary

Illumination invariance remains the most researched, yet the most challenging aspect of automatic face recognition. In this paper, the authors investigate the discriminative power of colour-based invariants in the presence of large illumination changes between training and test data, when appearance changes due to cast shadows and non-Lambertian effects are significant. Specifically, there are three main contributions: they employ a more sophisticated photometric model of the camera and show how its parameters can be estimated, they derive several novel colour-based face invariants, and on a large database of video sequences they examine and evaluate the largest number of colour-based representations in the literature.

  • Format: PDF
  • Size: 964.5 KB