Using advanced machine learning models, Google researchers have developed a new artificial intelligence system, called Vlogger.

Another AI technology is coming? Yes, and it can turn a photo into realistic videos. In addition, it makes it easy to create an AI influencer for social networks. Discover this promising technology now.

Creating realistic avatars

At the heart of this revolution, Google researchers have worked hard to bring Vlogger. This innovation has the power to transform a simple photograph into an animated avatar, commanded directly by the voice of its creator.

Even if Vlogger hasn’t officially launched yetthe demonstrations that are circulating offer us a breathtaking glimpse of realism.

It is true, of course, that other tools allowing similar animations were already present on the market, like the lip synchronization offered by Pika Labs or even Hey Gen video translation services And Synthesis.

However, Vlogger stands out radically thanks to its more intuitive approach. In addition, its reduced consumption in terms of resources signals a real revolution.

The operation of Vlogger is based on an advanced broadcast architecture. This technology goes through several steps to generate the avatar.

First, it analyzes the audio and image provided, submits them to a process of creating movement in 3D. Then, it uses a temporal diffusion model to determine movements and their timing.

Finally, the avatar is adjusted to produce the final result. The system relies on a neural network that predicts facial and body movement and facial expressions. He uses the still image as a starting point and the audio as a guide.

For training, the model used MENTOR, a large dataset containing labeled videos of people speaking. This greatly enriched its ability to generate realistic avatars.

Exploring potential and recognizing limits

Despite its considerable potential, it should be recognized that Vlogger is not without limitations. Currently in the state of a prototype rather than a finished product, it may does not reproduce natural movements with absolute fidelity of an individual.

The researchers themselves admit that, faced with complex gestures or in heterogeneous environments, the model could be less efficient. It is, moreover, optimized mainly for the creation of short video sequences.

However, the table is not only tinged with dark shades, Vlogger’s potential fields of application are, in fact, particularly extensive. They extend, on the one hand, to improving video translations and to designing virtual assistants convincingly animated. On the other hand, they include the development of video game characters with striking realism.

A particularly exciting point is the opportunity to exploit Vlogger. Indeed, this makes it possible to optimize video communications, especially in low bandwidth contexts. This perspective opens the door to much more fluid and accessible virtual exchanges.

    Share the article:

Our blog is powered by readers. When you purchase through links on our site, we may earn an affiliate commission.

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *