Google’s VLOGGER AI Can Create Realistic Video Avatars From A Single Image: Here’s What You Need To Know
Alphabet Inc. (NASDAQ:GOOG) (NASDAQ:GOOGL)-owned Google AI team’s latest project that can generate high-resolution video avatars from a single image is all the rage on social media nowadays.
Here’s everything you need to know about this development in the AI arena.
What Happened: Earlier this month, Google’s Enric Corona and his team introduced VLOGGER, a tool that can produce high-quality video avatars from a single image. These avatars can be animated to mimic a person’s speech, hand and body movement, blinking, and facial expressions, creating a highly realistic representation.
How Does It Work: VLOGGER employs a two-stage pipeline to produce lifelike talking videos. The first stage derives body motion commands from audio input, encompassing gaze, facial expressions, and posture. These commands are used to generate frames through image-to-image translation in the second stage, with consideration of a reference image for identity conditioning.
Subscribe to the Benzinga Tech Trends newsletter to get all the latest tech developments delivered to your inbox.
What Are Its Applications? According to Google researchers, this model’s primary application lies in video translation. For example, VLOGGER can process an existing video in a specific language, adjusting the lip and facial regions to synchronize with new audio inputs.
Other potential applications include creating animated avatars suitable for virtual assistants, chatbots, or virtual characters exhibiting realistic appearances and movements within gaming environments.
When Will It Be Available To Masses: At this point, VLOGGER exists solely as a research project with a few entertaining demo videos. It is not available to try. However, the demo videos have already given people plenty of ideas that can be implemented if they get turned into a product.
