Mitigation of Rotational Constraints in Image Based Plagiarism Detection Using Perceptual Hash
Sushila Palwe, Ankit Kumar, Ganesh Gadakar, Aditya Mehta, Amey Mhaisane · 2014
Image plagiarism can be detected using similarity between the two images. The perceptual hash function calculates similar hash value for similar images. This paper proposes a new modification to perceptual hashing (pHash) based approach for Image Retrieval. By the simplified use of a distance function two perceptual hash values can be compared and, it can be understood whether the images being compared are perceptually different or not. Existing Image Plagiarism detection techniques are not sensitive to image rotation. We propose an algorithm that compares the various degrees of rotation of the source image to the perceived similar image in an efficient way. Perceptual image hash functions along with the rotation check can be used for the identification of plagiarised images, authentication or integrity verification of images.