An approach for detection of forgeries in digital images

Abstract

The essential aim of the work reported in this thesis is to upgrade the current degree of comprehension of the subject of image content verification and present practical tamper detection strategies and solutions that remain apposite to a wide scope of pragmatic forensic situations. newlineThe outcomes of broad experimentation validate the effectiveness of the proposed arrangements and further demonstrate that the proposed approach is fit for fulfilling the necessities of true real-world forensic situations. Broadly speaking, the proposed approach was able to detect forgeries with the precision rates of 94.59% and 99.49%, and accuracy of 93.86% and 96.79% on benchmark datasets viz. CASIA v1 and CASIA v2, respectively. newlineAnother approach is proposed to improve the detection accuracy by using the combination of two streams of features to form a more distinctive representation apt for content authentication of digital images. Consequently, by joining the two categories of image features, the detection accuracy is fundamentally upgraded compared to using a single method and other contemporary methods. The proposed fusion-based approach used 1160-dimensional features and has state-of-art accuracy of 97.94% and 99.31% on CASIA v1 and CASIA v2 datasets, respectively. The proposed approach is promising for offline forensic analysis of digital images. However, for real-time analysis, the high dimensionality of fused features is the primary bottleneck. newline newline