Detecting image manipulation with ELA-CNN integration: a powerful framework for authenticity verification.

Area of Science:

• Computer Vision
• Digital Image Forensics
• Machine Learning

Background:

• Rapid advancements in image editing software have led to an increase in sophisticated image forgeries.
• Existing image manipulation detection techniques require further improvements in accuracy and precision.

Purpose of the Study:

• To propose a novel algorithm for detecting image forgeries, specifically copy-move and splicing attacks.
• To enhance the accuracy and reliability of digital image forensics.

Main Methods:

• The proposed algorithm integrates Error Level Analysis (ELA) with a Convolutional Neural Network (CNN).
• ELA identifies regions with varying compression levels, and these ELA images are used to train a CNN model.
• The CNN architecture comprises convolution, max pooling, dense layers, and dropout for generalization.

Main Results:

• The algorithm achieved a training accuracy of 99.05% and a testing accuracy of 94.14% on the CASIA 2 dataset.
• The system demonstrated high precision (94.1%) and recall (94.07%) in detecting image forgeries.
• The proposed method outperformed state-of-the-art techniques in both accuracy and precision.

Conclusions:

• The integration of ELA and CNN offers a robust and effective solution for image forgery detection.
• This approach shows significant promise in combating the proliferation of deceptive visual content.
• Further research can build upon this method to address increasingly complex image manipulation techniques.