Author(s): D Shine Rajesh, S Vyshnavi, B Pranavi and J Sravani

Abstract: The development of more sophisticated tools for manipulating digital images has spurred renewed interest in the topic of digital picture fraud detection. This work delves into the topic of Copy Move Forgery Detection (CMFD), a method for passively identifying pictures that have been altered using the copy move technique. This paper proposes a convolutional neural network (CMFD) method for feature extraction and feature matching using two different algorithms: oriented features from the Accelerated Segment Test and rotated binary robust independent elementary features (Oriented FAST and rotated BRIEF, respectively). Images that were subjected to different geometrical assaults were used to evaluate the proposed CMFD approach. When tested using pictures from the MICC-F600 and MICC-F2000 databases, the suggested method yields an overall accuracy rate of 84.33% and 82.79%, respectively. With object translation, varying degrees of rotation, and magnification, forgery detection was able to reach a True Positive Rate of over 91% for altered photos.

DOI: 10.33545/27076571.2024.v5.i2b.99

Pages: 95-98 | Views: 594 | Downloads: 209

Download Full Article: Click Here