Deepfake Detection using Inception-ResnetV2

Deep learning has benefited us in resolving many complex problems. Computer vision is a subcategory of it. With the ability to find patterns from unstructured data, Deep learning has immense potential. Big techs are very keen on producing a computer with human brain-like decision-making capabilities. With all these sweeter sides comes the bitter side of it. Deepfake is one such occurrence. It creates a mask which contain properties of a particular person and can be applied to some other person. In this way the target is depicted doing deeds which he never did. With the increased capacity of a specific field i.e. Generative Adversarial Network (GAN); now we can create high-quality deepfakes. Deepfakes nowadays can easily deceive human eyes. The consequences of this can be devastating and unforeseeable. Creating chaos, privacy threats are some of the major reasons why people are questioning deepfakes. Victims’ size has started including common public. What could be done is to keep a check over its spreading. This work has taken into consideration the problems that emerged by deepfakes and proposed a method to detect forgery among videos.