Comparison of RADIAL variance based and Maar-Hildreth operator perceptual image hash functions on biometric templates

This paper presents a practical comparison between two perceptual hash functions, RADIAL Variance Based Hash and Maar-Hildreth Operator. Perceptual image hashing functions are used for image and visual content authentication. For the purposes of this paper the authors will compare the two functions to evaluate their resistance to several types of attacks on biometric templates.  The main objective of this paper is to examine whether the implementation of perceptive hash functions could have an effect on similarity of these templates. pHash, an open source implementation of various perceptual hash functions, was used to benchmark these functions. The results of this comparison and evaluation will be preliminary because only three most common biometrics characteristic templates will be considered. Robust perceptual image hashing methods will be used to overcome the abovementioned problems of a challenging developing biometric systems.