Perceptual hash (PH) algorithms are image hashing algorithms that output visual content-based hashes. Unlike cryptographic hashes, where a minute difference in input data has an enormous difference in its hashes, PH algorithms change proportionately to the change in their inputs, i.e., trivial differences between two images result in insignificant differences between their hashes.PH algorithm’s application over the years has spiked; for example, in digital forgery detection, data authentication, and preventing nonconsensual data propagation. This study compares several well-known and publicly available hash algorithms for biometric fingerprint authentication, including Average Hash, Discrete Cosine Transform (DCT) based Hash, Difference Hash, and Wavelet-based Hash. To evaluate the performance of these algorithms, we consider critical factors like discrimination capability and robustness to variations and distortions applied on a dataset of biometric fingerprints. The outcome of this study will be the best hash algorithm suitable for biometric fingerprints. Our work is to initiate a pipeline to investigate biometric authentication using PH algorithms. Key Words: Cryptographic Hashes, Perceptual Hashing, Biometric Fingerprints.

Recommended Citation: Nagendrappa R., Jafarikhah S., Stoker G., Layman L., Siadati S., (2023). Evaluation of Perceptual Hashing Algorithms on Biometric Fingerprints. UNCW MS CSIS Proceedings. V. 17 , N. 10 .