Copyright © 2021 Euromaidanpress.com

The work of Euromaidan Press is supported by the International Renaissance Foundation

When referencing our materials, please include an active hyperlink to the Euromaidan Press material and a maximum 500-character extract of the story. To reprint anything longer, written permission must be acquired from [email protected].

Privacy and Cookie Policies.

Ukraine startup RefaceAI targets global market with new deepfake technology

Ukraine startup RefaceAI targets global market with new deepfake technology
Article by:  Mike Buryk, Ukraine Digital News

Thus, Ukrainian startup Looksery – founded in 2013 – developed an application that does facial modification of photos in realtime on mobile platforms. The company was acquired by Snapchat in 2015.

Today, one of the newest approaches in digital image modification is artificial intelligence with generative adversarial networks (GAN). A Ukrainian startup, RefaceAI, utilizes this technique, which is based on machine learning.

Unlike other applications such as Zao that rely on a 3D model of your face, RefaceAI’s Doublicat app allows you to superimpose your face onto any GIF in seconds.

[youtube https://www.youtube.com/watch?v=IQ7pn-u0gOs?feature=oembed]

The ancestor of RefaceAI was first established in 2011 by Oles Petriv, Denys Dmitrenko, Roman Mogylnyi, and Yaroslav Boiko. The company was called Neocortext. At first, they focused on using machine learning for sentiment analysis in texts, creating a system of language dynamics analysis and multilingual language tone analysis, but by 2018 the company took a different turn with their new software called Reflect Faceswap based on AI-face replacement technology.

In 2019, Dmytro Shvets, who had been a managing partner of a successful seed-stage venture capital fund, was invited by the team to become a co-founder of the company – rebranded to RefaceAI – with the title chief business officer.

“We use state-of-the-art GAN technology to swap faces in GIFs, digital images, and videos,” says Shvets. He points out that there is a practical application for the technology in many industries including entertainment, stock photo, gaming, and advertising.

Because their product can do up to 400,000 face-swaps per hour, Shvets says that it can also be utilized in streaming video.

While the 40-person development team for RefaceAI is based in Kyiv (Kiev), the company is actually incorporated in the State of Delaware.

The plan is to eventually open an office in the US for marketing and sales since many of their clients are located there.  Both Los Angeles and Silicon Valley are top targets for future marketing efforts.

RefaceAI has received media coverage in MashableForbesThe Verge, and Digital Trends.

Face-swapping technology is not without its controversy. In early 2020, Facebook amended its user policies to ban some types of so-called deepfake videos that could be used to manipulate voters in the upcoming election.

“Together with more advanced versions and broader access to our face-swap technology, we’ll launch a deepfake detection tool,” says Shvets.

  • You can listen to a recent podcast interview with Dmitry Shvets here
You could close this page. Or you could join our community and help us produce more materials like this.  We keep our reporting open and accessible to everyone because we believe in the power of free information. This is why our small, cost-effective team depends on the support of readers like you to bring deliver timely news, quality analysis, and on-the-ground reports about Russia's war against Ukraine and Ukraine's struggle to build a democratic society. A little bit goes a long way: for as little as the cost of one cup of coffee a month, you can help build bridges between Ukraine and the rest of the world, plus become a co-creator and vote for topics we should cover next. Become a patron or see other ways to support. Become a Patron!

To suggest a correction or clarification, write to us here

You can also highlight the text and press Ctrl + Enter

Please leave your suggestions or corrections here