Robust illumination-invariant features by quantitative bilateral symmetry detection

In this paper we propose a novel method for the determination of illumination-invariant features in images. Quantitative bilateral symmetry of a given scene is computed using dynamic programming. In contrast to problems with other methods, the results of the dynamic programming algorithm describe symmetry in terms of an absolute region instead of a relative degree. Vertical symmetry images are generated and symmetry axes are extracted using non-maxima suppression and hysteresis thresholding. For each symmetry image a unique feature vector is obtained as the sum of the gray-values of each column of the image. The usefulness of the proposed feature detection algorithm is shown by a preliminary experiment. In the experiment, the feature vector is used to robustly track motion in an image sequence.