Kevin R. Hudson

An image-clustering method based on cross-correlation of color histograms

Color histogram analysis is a powerful tool for characterizing color images, and has been widely used in image indexing and retrieval systems. Image classification using histogram analysis is possible if a robust similarity measure between different color histograms can be found. In this paper, we propose to use a cross-correlation function to measure color histogram similarity. We will show that a cross-correlation function has several advantages over the method of histogram intersection, which has been widely used to calculate the similarity between color histograms. The cross-correlation function is normalized automatically, determines similarity irrespective of image size, is invariant to small color shifts, and is easier to implement using computationally efficient methods. We present an example of unsupervised image clustering by applying a cross-correlation function to color histograms. This method was used to improve the perceived color consistency in a multi-print-engine system. We also show how to optimize the cross-correlation function to compensate for color shifts.