Abstract:The proliferation of manipulation of images and videos and its resulting distortion of individuals or impact on public opinion in general is due to the ease of obtaining images and videos for their spread on social media platforms. Furthermore, the ease of using artificial intelligence programs and editing them with high resolution makes it difficult for the human eye to distinguish between them. Manipulated digital photos and videos play a crucial role in digital criminal investigations, such as the dissemination of fake multimedia and digital kidnapping-related crimes. Automation of the investigation process is an effective approach that is widely utilized in digital criminal investigations. For example, digital forensic analysis tools are being used to identify digital evidence in cybercrimes. Due to the complexity of the analysis of these crimes, there is a need to implement efficient digital forensics approaches based on recent technologies. Therefore, automating the detection of fake images and videos is challenging. According to state-of-the-art techniques, Machine Learning (ML) and Deep Learning (DL) methods can play a significant role in mitigating this challenge. In this paper, several machine learning methods, such as Support Vector Machine (SVM), K-Nearest Neighbor (KNN), and Wide Neural Networks (WNN), were applied to the Saraferreirascf/Photos-Videos-Manipulations- dataset, which includes 40,588 images and 12,400 images extracted from video (real and fake). The findings of this study show that SVM with a Bayesian optimizer obtained the highest accuracy of 89.2% for videos, significantly outperforming other machine learning methods used in this paper. It also provided the highest accuracy of 99.47% for images, although with slight differences compared to other machine learning methods. Moreover, SVM with a Bayesian optimizer achieved an accuracy of 94.68% for both videos and images together.

A Novel Machine Learning Approach for Forgery Detection and Verification in Digital Image

An Evaluation of Digital Image Forgery Detection Approaches

Comprehensive analyses of image forgery detection methods from traditional to deep learning approaches: an evaluation

Image Forgery Detection: Survey and Future Directions

Detecting and Localizing Copy-Move and Image-Splicing Forgery

Tampering Detection System ( A Machine Learning Approach )

A deep learning framework for copy-move forgery detection in digital images

An efficient approach for copy-move image forgery detection using convolution neural network

A survey on digital image forensic methods based on blind forgery detection

Design of Automated Deep Learning-Based Fusion Model for Copy-Move Image Forgery Detection

A Novel Approach for Detecting Forged Image.

A Review on Copy-Move Image Forgery Detection Techniques

Analysing Statistical methods for Automatic Detection of Image Forgery

Machine learning-based approaches for manipulated image and video forensics in digital criminal investigation

Passive Image Forgery Detection Techniques: A Review, Challenges, and Future Directions

Image forgery detection: a survey of recent deep-learning approaches

Detecting video forgery: A machine learning approach for consistency analysis of video frames

A Comprehensive Review of Deep-Learning-Based Methods for Image Forensics

Image Forgery Detection and Localization Using Block Based and Key-Point Based Feature Matching Forensic Investigation

Image Inpainting Forgery Detection: A Review

Detection of Copy-Move Forgery in Digital Image Using Multi-scale, Multi-stage Deep Learning Model