Embedded Hash Codes for Image Similarity Detection and Tamper Proofing

Abstract

The image and video traffic on the Internet has increased due to the ease and abundance of the image capture and sharing services. However, this can sometimes result in violations of copyright and image tampering for various purposes. Finding differences between images has importance for image retrieval, aiding forensics, disease diagnosis, and environmental changes etc. Many similarity measures have been proposed that use different features extractable from the detected image. In some cases it is important to determine if the set of images differ and if so then to what extent. Digital Rights Management (DRM) has gained prominence and importance due to widespread availability of tools to modify images. Sometimes the need arises to claim the rightful ownership of an image. The blockchain through the use of hash codes can be used to determine tampering similar to its success in many industries where privacy and security of the documents is important. In this paper, we embed the image hash code in the JPEG image header. The proposed system simplifies the hashing and encryption process by processing only the image data without which a reconstruction is not possible. The results show that the image data can thus be shared without the risk of any undetected tampering or misuse.