Detection of Fake Videos using Convolutional Generative Method
Keywords:Fake Vidoes, GAN, Supervised Learning, Unsupervised Learning, DCGAN
Fake videos are used in different industries for positive aspects, but mostly, people use fake videos to defame politicians and celebrities. Fake videos create great social and security concerns because people use fake videos and images to gain illegal access to biometric security systems. Detection of fake videos is a challenging task. Recently deep learning methods have been applied to solve this problem. A generative novel deep convolutional generative adversarial network (DCGAN) is proposed to detect fake videos in this research work. The proposed novel model is evaluated on celeb-DF and DFDC datasets with different batch and epoch sizes in this research work. The proposed novel DCGAN model gained the highest accuracy of 96% on a celeb-DF dataset, and the DFDC dataset gained an accuracy of 93.5%. The model is compared to available state-of-the-art methods.
How to Cite
This is an open Access Article published by Research Center of Computing & Biomedical Informatics (RCBI), Lahore, Pakistan under CCBY 4.0 International License