The AI research team at Samsung Electronics and Imperial College London developed a technology that creates a video that a person speaks with only one picture and audio data.
Deepfake refers to creating a fake video by combining the video with the existing video by learning a large amount of training data through machine learning. Machine learning technology was also used in this study. The deep fake created with one photo is not 100% realistic, but the completeness is amazing.
For example, it is possible to create a lecture video that has not actually been made by overlapping a photograph of Albert Einstein with the audio of the lecture. You can even have a Rasputin video sing something like Beyonce’s Halo. It also adds emotional expression.
Looking at this, it can be said that it is only a matter of time before the emergence of software that allows the general public to easily create deep fakes as deep fake technology is developing at a frightening rate. Related information can be found here .