Hybrid deep learning and machine learning approach for detecting spatial and temporal forgeries in videos

Video forgery detection has become increasingly critical due to the rise of sophisticated video manipulation techniques. Traditional methods often struggle to keep up with the ever-evolving sophistication of forgery techniques, necessitating innovative and advanced solutions. In response to this challenge, we propose a hybrid approach that combines deep learning and machine learning techniques. This approach leverages the power of DenseNet121 for excellent spatial feature extraction, utilizes a spatiotemporal autoencoder to capture spatiotemporal dependencies, and employs a gradient boosting machine (GBM) to effectively classify the video clips based on the extracted features. This provides a robust and comprehensive solution for detecting spatial and temporal forgeries in videos. By combining these components, this approach offers a more robust and comprehensive solution for detecting spatial and temporal forgeries in videos. To evaluate our approach, we are using a diverse dataset of videos (VTD and VIFFD) with different forgery scenarios and conducting extensive experiments. The results demonstrate the effectiveness and accuracy of our hybrid approach in detecting spatial and temporal forgeries, surpassing the performance of individual classifiers and traditional methods.