Real-Time Simulation of the Occlusion-Noise Attack and a Mitigation Proposal

Abstract

The performance of image transmission can be affected by noises. These can be produced by natural conditions or attacks induced by third parties. The main inconvenience of noise in images is the loss of visual quality. The higher the percentage of noise in the image, the lower the clarity of information obtained. For this reason, in this paper, occlusion noise is simulated in real-time during the download of an image. The simulation includes the malicious interference of a third agent with the data transfer between two mobile devices. In this way, the receiver realizes the damage only after the image reproduction is complete. The purpose is to demonstrate the effects of this type of noise and the difficulties in understanding damaged information. Additionally, the present work shows the utility of mitigation in occlusion attacks, despite sometimes not being completely avoided. The proposed solution involves a permutation that spreads the noise across all rows. Consequently, each row has both initial values and damaged pixels.