imPlag: Detecting image plagiarism using hierarchical near duplicate retrieval

Plagiarism in any form is a serious offense especially in academia and industry where integrity and royalty from work is of utmost importance. In this work, a novel hierarchical feature extraction as well as an approximate nearest neighbor search is proposed for detecting plagiarism of images. The proposed scheme is applicable for natural images as opposed to specific image classes reported in a previous work. A comprehensive experimental analysis is provided to illustrate the efficacy of the techniques chosen for the scheme. We demonstrate that the scheme shows a lot of promise for a wide variety of attacks and is amenable to scaling.