Runway’s stunning AI tool for simulating facial expressions + video

Artificial intelligence startup Runway has a new feature called “Act-One» has introduced that it allows users to record a video with their camera and then display people’s facial expressions very accurately in the form of artificial intelligence-generated animated characters.

According to Runway’s website, this free tool will gradually be available to users from today; This feature is labeled “coming soon” for us. The following video was published to introduce this feature, which shows its extraordinary performance in reconstructing facial expressions.

Stunning demo of Runway’s new AI tool

Anyone with a Runway account can use the feature, but must have enough credit to produce a video with the “Gen-3 Alpha” model that was introduced earlier this year.

Although this feature is currently available in a limited way, it has been very well received in social networks. Allen.T, which specializes in AI-powered video production, described Runway’s new tool as “a trump card.”

Cristobal Valenzuela, Runway’s co-founder and CEO, told VentureBeat that the feature currently focuses more on facial expressions. However, maybe in the future Runway will pay attention to the reconstruction of body movements.

AI models for video production, which were first released in late 2022, have since made great strides in terms of realism, clarity, and fidelity to user commands or descriptions. One of the areas that is still limited is the realistic drawing of facial expressions.

With Act-One, Runway aims to simplify the complex process of displaying facial expressions in artificial intelligence videos. With the help of this new tool, creators can produce anime characters in different styles and designs, and without the need for motion capture tools, they can also create completely realistic facial expressions. With this in mind, Act-One can have many benefits for animators, game developers, and filmmakers.