DeepFake Creation and Detection Using LSTM, ResNext

Technology was created as a means to make our lives easier. There is nothing more fast-paced than the advancements in the field of technology. Decades ago, virtual assistants were only a far-fetched imagination; now, these fantasies have become a reality. Machines have started to recognize speech and predict stock prices. Witnessing self-driving cars in the near future will be an anticipated wonderment. The underlying technology behind all these products is machine learning. Machine learning is ingrained in our lives in ways we cannot fathom. It may have many good sides but it is misused for personal and base motives. For example, various forged videos, images, and other content termed as DeepFakes are getting viral in a matter of seconds. Such videos and images can now be created with the usage of deep learning technology, which is a subset of machine learning. This article discusses the mechanism behind the creation and detection of DeepFakes. DeepFakes is a term generated from deep learning and fake. As the name suggests, it is the creation of fabricated and fake content, distributed in the form of videos and images. Deep learning is one of the burgeoning fields which has helped us to solve many intricate problems. It has been applied to fields like computer vision, natural processing language, and human-level control. However, in recent years, deep learning-based software has accelerated the creation of DeepFake videos and images without leaving any traces of falsification which can engender threats to privacy, democracy, and national security. The motivation behind this research article was to spread awareness among the digitally influenced youth of the twenty-first century about the amount of fabricated content that is circulated on the internet. This research article presents one algorithm used to create DeepFake videos and, more significantly, the detection of DeepFake videos by recapitulating the results of proposed methods. In addition, we also have discussed the positive aspects of DeepFake creation and detection, where they can be used and prove to be beneficial without causing any harm.