Morphological correlation for color image matching

Image matching is an area of intensive research. Among others, correlation methods have been used widely for this purpose. In literature several correlation filters have been proposed for image matching. Traditionally linear correlation is applied among the images, however the operation is not robust when images are corrupted with non-Gaussian noise. On the other hand, many recognition systems are based on grayscale images, although color images provide more information. In this paper we propose the use of morphological correlation combined with nonlinear filters for robust color image recognition. Traditionally, this scheme required threshold decomposition of grayscale images and correlation of each pair of binary slices. In order to reduce the great amount of binary correlations needed, we derive an expression to compute a single correlation between the grayscale images Performance of the proposed technique is compared with that of linear and nonlinear filters in terms of discrimination capability. Computer simulation results are provided and discussed.