HomeTech NewsAppsMeta presents an Artificial Intelligence capable of generating videos

Meta presents an Artificial Intelligence capable of generating videos

Published on

- Advertisement -
Share on Facebook Share on LinkedIn
- Advertisement -
Share on Pinterest Share to Email Share on Telegram Share on WhatsApp
- Advertisement -

Meta, parent company of Facebook, Instagram, WhatsApp, Oculus… has announced Make-A-Video, its own artificial intelligence (AI) capable of generating videos from a text description made in natural language.

The user dictates some instructions through natural text (prompt) that Make-A-Video transforms into a video

This is an important advance over other recently released AIs capable of generating static images from natural text descriptions (“prompts”). In this case they are moving images, animations that at the moment are of short duration but that already allow us to glimpse a future in which the creativity of users and their imagination are capable of being captured in photorealistic videos. Sizes, postures, positions, aspects… any element of the creation in the video can be modified to better suit what was in the imagination of the creator of that new content.

The user only needs to make a concise description in natural text of what he wants to appear on the screen and, as in other AIs, Make-A-Video shows different variants that, in addition, can be selected to add improvements with which to adjust even more to the result that the user wants.

- Advertisement -

Additionally Make-A-Video can from very basic imagesjust a few strokes, with which to mark basic parameters regarding the structure and elements of the video.

From Meta they recognize that despite the amazing results it is still a developing model and that it still needs to be perfected. To obtain the final images, large databases of public images are used while the results obtained are analyzed to train the AI ​​in such a way that it achieves the best results.

In addition to starting only prompts (text-only descriptions), Make-A-Video too allows you to include strokes or even images that serve as the basis for the generation of the videos, although the essence of this type of AI is the ability to explore the own imagination of the person who writes the prompt to, taking it as a basis, generate the video and offer a visual world that until then only existed in the world of ideas.

From Meta they raise the possibility of turning a project like Make-A-Video into a way of artistic expression that would have a special impact on the metaverse by allowing Meta environments to be created from the imagination of its users.

- Advertisement -

Latest articles

Croatia set to join Schengen but Austria ‘main hold-out’ against Romania and Bulgaria’s joint bid

The enlargement of the passport-free Schengen Area has been put under serious question ahead...

Nokia X10 and X20 start receiving Android 13 update

Update (12/07/2022) - by DT THE nokia started releasing this week the update with the...

More like this