Russian Developer Enhances Deepfake Technology

A Russian programmer has improved Deepfake technology by developing an algorithm that modifies not just the face, but the entire “persona.” According to the developer, all that’s needed is a single source photo, a target image, five approximate body parameters, and “0.1 seconds of patience” from the user.

After a process called “3Dfying”—which creates a three-dimensional version of a two-dimensional image—the result is a completely transformed persona. This includes the head, hair, skin color, and body shape, all of which replicate those in the original photo.

Unlike the older Deepfake technology, this new version changes not only the face but also the head, hair, skin tone, and even the figure. Moreover, the technology can alter any photo, not just those preloaded into the algorithm.

Developer Alexey Chaplygin claims he worked alone on this technology for two years, stating, “It was quite difficult to create something that no one else had managed before, especially when working solo.” Currently, the program runs on his “home computer,” but to scale the project, he needs “servers, high-speed internet, and a team to support the entire setup.” Chaplygin plans to seek funding for mass implementation from interested individuals or organizations through the “six degrees of separation” method.

The developer admits that such “deepfakes” could potentially be used for “identity theft or forgery,” but insists that Deepfake 2.0 was created for “good purposes.” One intended use is a virtual fitting room for online clothing stores, where anyone can upload their photo to see how a future purchase might look.

Deepfake 2.0 can be integrated as a plugin into any software or device, allowing users to instantly see themselves in any photo. The technology can also be used in photo editors, film production, animation, and more.

The quality of the new algorithm can be seen in a demonstration video published on Alexey Chaplygin’s YouTube channel.