Image Forgery Detection Using Deep Learning: A Comparative Study

Document Type : Original Article

Authors

1 Computer Science Dept., Faculty of Computers and Artificial Intelligence, Helwan University, Cairo, Egypt

2 Department of Computer Engineering MTC

Abstract

In recent years, there has been a significant increase in online activities, including business meetings, education, research, and virtual conferences. As a result, digital images have become the main source of information that can be shared and visualized on social media, in addition, it’s easy to forge these images using image-editing software, and it’s essential to detect image forgery for such images. So, it becomes essential to introduce an efficient image forgery detection technique to classify these images as either authentic or forged. In the past few years, deep learning-based techniques have achieved remarkable results in the field of image forgery detection IFD, most of them used transfer learning with the help of pre-trained models aiming to reduce time in the training and detection phase. This paper presents a comparative study of various image forgery detection techniques, it explores the techniques based on new deep learning models and techniques based on transfer learning models with the help of pre-trained models. The study aims to provide insights into the performance of different techniques used in deep learning and pre-trained models in image forgery detection, which may guide any researcher to present a useful model, that can detect multiple image forgery types simultaneously with improved detection accuracy and minimal detection time. The discussed results suggest the use of pre-trained models in the feature extraction phase only. and recommend using deep learning in classification.

Keywords

Main Subjects

Files

Share

How to cite

Statistics