VIOLA jones algorithm with capsule graph network for deepfake detection

Area of Science:

• Computer Science
• Artificial Intelligence
• Machine Learning

Background:

• DeepFake technology poses a significant challenge due to its sophisticated forgery capabilities.
• Current DeepFake detection methods struggle with accuracy and are limited in scope.
• Advancements in deep learning necessitate improved and intelligent fake detection algorithms.

Purpose of the Study:

• To propose an advanced deep learning model for robust DeepFake detection.
• To enhance the feature selection and extraction process for improved detection accuracy.
• To address the limitations of existing fake content detection techniques.

Main Methods:

• Utilized the Viola-Jones (VJ) algorithm for optimal feature selection.
• Implemented a Capsule Graph Neural Network (CN) for enhanced fake detection.
• Improved the graph neural network by incorporating capsule-based node feature extraction.

Main Results:

• The proposed model demonstrated a significant improvement in DeepFake detection accuracy.
• Experiments were conducted on combined CelebDF-FaceForensics++ (c23) datasets.
• The model achieved a high accuracy rate of 94% in detecting forged content.

Conclusions:

• The developed deep learning model offers a more effective solution for DeepFake detection.
• Capsule Graph Neural Networks show promise in advancing multimedia forensics.
• The research contributes to the ongoing effort to combat sophisticated digital manipulation.