Adversarially robust deepfake media detection using fused convolutional neural network predictions.

摘要: Deepfakes are synthetically generated images, videos or audios, which fraudsters use to manipulate legitimate information. Current deepfake detection systems struggle against unseen data. To address this, we employ three different deep Convolutional Neural Network (CNN) models, (1) VGG16, (2) InceptionV3, and (3) XceptionNet classify fake real images extracted from videos. We also constructed a fusion of the CNN models improve robustness generalisation capability. The proposed technique outperforms state-of-the-art with 96.5% accuracy, when tested on publicly available DeepFake Detection Challenge (DFDC) test data, comprising 400 model achieves 99% accuracy lower quality DeepFake-TIMIT dataset 91.88% higher In addition prove that prediction is more robust adversarial attacks. If one compromised by an attack, does not let it affect overall classification.