Closed Frequent Itemset Mining Approach to Image Security Enhancement

Abstract

With the rapid development of multimedia applications which involves images, video and audio, security has become an important aspect of modern day digital communication. Various fields like multimedia systems, telemedicine, military communications, medical imaging internet communications etc, are widely based on image security. Our focus in this paper is to propose an algorithm which enhances image security. The algorithm proposed provides two-level security for the images that are transmitted over networks. We are using an existing and novel Data Mining technique, called Closed Frequent Itemset Mining to encode the image. Upon this, we are applying an image security technique called Steganography to hide the very existence of the image whilst being sent. With this higher security levels, an image can be sent securely over any network with the image's existence completely hidden. Two prominent quality metrics are tested: They are Mean Square Error (MSE) and Peak Signal to Noise Ratio (PSNR). Mean Square Error has been reduced efficiently and Peak Signal to Noise Ratio showed much improvement, which proves that the algorithm not only provides multi-layer security but also preserves the quality of images. The effect of Minimum Support Count and message image size on PSNR has also been observed. Also, their effect on time taken by the algorithm has been observed and plotted.