Content Based Image Hashing Robust to Geometric Transformations

An image hashing provides a compact representation of an image that can be used for authenticating that image. This paper presents a new approach for generating an image hashing that is resilient to content-preserving modifications and at the same time, is capable of detecting malicious tampering. The invariant features are extracted via DWT and Radon transform followed by log mapping and then Fourier transform. In detection, the normalized Hamming distance is employed to weigh the similarity of the image hashing. Experimental results show that the proposed method can resist insignificant modifications such as compression, moderate geometric and filtering distortions. It is also sensitive to content changing (malicious) manipulations of image data.