Vol. 10 No. 2 (2023): Vol 10, Iss 2, Year 2023
Articles

FSM and RRU-net module for Image Splicing Forgery Detection

Velmurugan S
Department of Computer Science, Kongunadu Arts and Science College, Coimbatore -641029, Tamil Nadu, India
Saravana Moorthy R
Department of Computer Science, Kongunadu Arts and Science College, Coimbatore -641029, Tamil Nadu, India
Subramanian K
Department of IT and Analytics, Xavier Institute of Management and Entrepreneurship (XIME), Bangalore, India
Angel S
Department of Computer Science (SF), Avanishilingam University, Coimbatore, Tamil Nadu, India
Published December 31, 2023
Keywords
  • RU-Net, residual propagation, residual feedback, spliced image
How to Cite
S, V., R, S. M., K, S., & S, A. (2023). FSM and RRU-net module for Image Splicing Forgery Detection. Kongunadu Research Journal, 10(2), 81 - 88. https://doi.org/10.26524/krj.2023.19

Abstract

Nowadays, it can be difficult to tell whether an image is real or fake. Thanks to technological advancements, an image can be altered or falsified in a matter of seconds. Finding these forgeries has grown to be a major problem in the modern world. Although an image could be crucial evidence, it will be useless if it is faked. Methods for distinguishing between pictures that have been edited and those that have been computer-generated must be developed. In order to identify these forgeries, we plan to create an Image Forgery Detection Model that combines FSM and RRU-Net. Residual propagation and residual feedback are two distinct approaches that are combined in RRU-Net, which stands for Ringed Residual Structure and Network Architecture. To find long-distance dependencies, the Feature Similarity Module, or FSM, will be employed. Our suggested system combines FSM and RRU-Net to improve accuracy. We will extract the differences in the picture attributes between the modified and unmodified parts using image patches of different sizes. Once the forged area has been identified, the final region will be shown in color. The method will prove useful in the future for identifying different types of spliced image frauds that appear on different social media platforms.

Downloads

Download data is not yet available.