Three-Dimensional Convolutional Approaches for the Verification of Deepfake Videos: The Effect of Image Depth Size on Authentication Performance

Deep learning has proven to be particularly effective in tasks such as data analysis, computer vision, and human control. However, as this method has become more advanced, it has also led to the creation of DeepFake video sequences and images in which alterations can be made without immediately appealing to the viewer. These technological advancements have introduced new security threats, including in the field of education. For example, in online exams and tests conducted through video conferencing, individuals may use Deepfake technology to impersonate another person, potentially allowing them to cheat by having someone else take the exam in their place. Several detection approaches have been proposed to address these issues, including systems that use both spatial and temporal features. However, existing approaches have limitations regarding detection accuracy and overall effectiveness. The paper proposes a technique for detecting Deepfakes that combines temporal analysis with convolutional neural networks. The study explores various 3-D Convolutional Neural Networks-based (CNN-based) model approaches and different sequence lengths of facial photos. The results indicate that using a 3-D CNN model with 16 sequential face images as input can detect Deepfakes with up to 97.3 percent accuracy on the FaceForensic dataset. Detecting Deepfakes is crucial as they pose a threat to the authenticity of visual media. The proposed technique offers a promising solution to this issue.