The Capacity and the Sensitivity of Color Histogram Indexing
Markus A. Stricker, Michael J. Swain · 1994
Color histogram matching has been shown to be a promising way of quickly indexing into a large image database. Yet, few experiments have been done to test the method on truly large databases, and even if they were performed, they would give little guidance to a user wondering if the technique would be useful with his or her database. In this paper we define and analyze measures relevant to extending color histogram indexing to large databases: capacity (how many distinguishable histograms can be stored) and sensitivity (how the average number of retrieved images depends on the retrieval threshold). The theoretical results lead to a practical test procedure which enables a user to determine the performance of color histogram indexing on a large database by looking at a small, randomly-chosen subset of the images. We suggest how our analysis can be extended to other feature-based indexing techniques. Capacity and Sensitivity of Color Histogram Indexing 1 1 Introduction As the cost of...