Kelsey A. Ramirez-Gutierrez, Mariko Nakano-Miyatake, and Hector M. Perez-Meana
Perceptual hashing, multimedia finger printing, Radon transform, image normalization
This paper proposes an improvement method using image normalization for Radon transform (RT)-based perceptual image hashing in which a normalized image is generated before the image hashing estimation doing it invariant to a wide range of geometric distortions. Proposed scheme improves the robustness of previously proposed RT-based image hashing schemes against geometric distortions. The proposed improvement is evaluated using two efficient RT-based image hashing algorithms, which are Wu’s and Seo’s algorithms. The evaluation results demonstrate that the proposed method, significantly improves the efficiency of Wu’s and Seo’s image hashing algorithms when the images to be analysed are distorted by several geometric attacks, such as rotation, scaling and shearing.
Important Links:
Go Back
