THE BRIDGE

tag Xpression Avatar

Gen AI startup EmbodyMe unveils new app, creates avatars responding to your motion in real time

SHARE:

Tokyo-based EmbodyMe launched a new app called Xpression Avatar, which allows users to move their own avatars in real-time, available for iOS and Android. The app uses the company’s proprietary real-time video generative AI technology to instantly generate your avatar in different styles which responds to your facial expressions and head movements fin real time. Since its launch back in 2016, EmbodyMe has developed several apps in the generative video field, including the EmbodyMe VR app, the Xpression Camera face swap app, as well as the Xpression Chat VR- / ChatGPT-based app. In the new app, the AI technology allows you to generate your avatar in various styles from your photos. The avatar styles available include anime, ukiyoe, humanoid, 1990s, and hip-hop. The user can not only move the avatar in accordance with his/her own movements, but also make the avatar speak his/her favorite lines, and tap a button to make the avatar laugh or sing. The generated images can also be shared on social network services while the mobile screen with the avatar can be shared. The new app is the culmination of EmbodyMe in two ways. One is that, in addition to the preset styles mentioned above, you…

Tokyo-based EmbodyMe launched a new app called Xpression Avatar, which allows users to move their own avatars in real-time, available for iOS and Android. The app uses the company’s proprietary real-time video generative AI technology to instantly generate your avatar in different styles which responds to your facial expressions and head movements fin real time.

Since its launch back in 2016, EmbodyMe has developed several apps in the generative video field, including the EmbodyMe VR app, the Xpression Camera face swap app, as well as the Xpression Chat VR- / ChatGPT-based app. In the new app, the AI technology allows you to generate your avatar in various styles from your photos.

The avatar styles available include anime, ukiyoe, humanoid, 1990s, and hip-hop. The user can not only move the avatar in accordance with his/her own movements, but also make the avatar speak his/her favorite lines, and tap a button to make the avatar laugh or sing. The generated images can also be shared on social network services while the mobile screen with the avatar can be shared.

Image credit: EmbodyMe

The new app is the culmination of EmbodyMe in two ways. One is that, in addition to the preset styles mentioned above, you can freely specify your favorite style with text prompts; according to EmbodyMe CEO Issay Yoshida, your prompt is interpreted based on the customized version of the Stable Difussion LLM (Large Language Model).

The second is their proprietary AI technology that can generate video in real time. Moreover, video processing is not performed in the cloud but on the mobile, which makes the company stand out with its unrivaled technology. The company does not need to increase its computational resources as its user base grows because the experience does not rely on the cloud.

EmbodyMe is developing the app for consumers while offering business solutions using the same technology for video production, advertising, live streaming, games, and other applications.

In the generative video AI space, Meta and Stable AI have recently announced their own solutions respectively while Runway has partnered with Getty Images to develop generative AI video models for the film and advertising industries. Pika Labs has reached a $200 million valuation in just six months after its launch.