GDP-PIH: A Robust Perceptual Image Hashing using Gradient Directional Pattern

Perceptual image hashing generates a compact representation of an image resistant to perceptual changes, such as those caused by geometric transformations. The resulting hash value can efficiently identify and authenticate images in large databases and detect image tampering or duplication. But, generating an efficient perceptual hash with good discrimination and robustness properties is challenging. This paper uses the Gradient Directional Pattern feature extraction technique to generate the robust perceptual hash. The proposed hash is experimented with single and multiple types of attacks, and yet it yields an efficient result. The experimentation results indicate that setting the threshold value to 0.25 achieves an optimal balance between the robustness and discrimination capabilities of the presented image hash.