Deep learning / Deepfakes Research Proessor Mahmood, Student Fahd M. Saleem (G01154081) IT-104-214 (Fall 2018)
The roots of deep fakes technology rely similarly on any other, according to “Phil's Stock World: Detecting 'deepfake' videos in the blink of an eye” the algorithm of deep fakes is similar to Googles translate service. Essentially AI uses machine learning to analyze hundreds to thousands of images of a person usually a celebrity. The images are synthesized onto another person’s face down to the various analogous movements & angles making it appear as the source is the target person (Stock, 2018). In essence, the more source information provided to the AI, the better the quality of the superimposed digital impression. Assumedly it can lead to many cons associated with the technology but stores excellent capabilities as well. Deep learning in addition to computer-generated imagery or more commonly known as CGI, offers a preeminent potential for film & television. London’s “The Economist” magazine by the title of “What if AI made actors immortal?: Performance anxiety” offers excellent insight into the potential uses for deep learning technology as a whole. Deep learning technology although already advance has the potential to carry on actors lives through film even after their death. Imagine an actor able to star in countless films at once. The magazine then goes on to say that different ages of the actors could be portrayed in various films. The real-world application of deep learning offers the legacy of beloved actors to live on for decades to come.
