Mathematical performance evaluation tool for image hash generation functions

Image hash as a descriptor or identifier can be used to measure similarity in images. Among all image hash functions, a histogram-based method shows better robustness to common signal processing operations and various geometric attacks. In this paper, performance of this method is improved by utilizing histogram generation method. Histogram with equal population improves the performance slightly. In order to compare performance of image hash generation functions, a statistical tool is provided. The proposed tool will help to evaluate the performance of image hash generator functions.