Meta launches AI-powered image and video editing tools

Meta says its new tool will allow users to edit images and generate videos for Facebook and Instagram.

Social media giant Meta has launched two new generative artificial intelligence (AI)-based tools for editing images and videos uploaded by Facebook and Instagram.

Meta said in a Nov. 16 post that Emu Video and Emu Edit will allow users to edit videos and images using text prompts. The tools are built on top of Meta's Emu, the company's first image generation base model.

The social media company further stated that the potential use cases for these tools are limitless as they can help people express themselves in new ways.

Meta did not disclose when the tools will be publicly available to users. The company has not yet responded to CryptoSlate's request for additional comment.

Emu Video

Emu Video allows users to create four-second long videos using text prompts and reference images. According to Meta, Emu Video leverages the company's Emu model as well as text-to-video capabilities based on generative models.

There are two steps involved in the video editing process. First, the user uses a text prompt to generate an image. They then use the previously generated image and its corresponding caption to create the video.

In addition, the tool can also "animate" user-provided images based on text prompts.

Meta says:

"In human evaluations, our video generation is strongly favored compared to previous work. In fact, 96% of respondents prefer the model based on quality, and 85% prefer it based on fidelity to text prompts".

Emu Edit

Emu Edit provides users with a user-friendly tool to adjust images easily.

According to the company, the tool "simplifies a wide range of image processing tasks and enhances the functionality and precision of image editing." ”

The tool will allow users to manipulate the background of the image, adjust the color and geometry of the objects in the image, and perform many other functions.

Meta says:

"Emu Edit follows the instructions precisely, ensuring that the pixels in the input image that are not related to the instructions remain intact. ”

Meta's Emu Edit tool is able to achieve this level of precision because it relies on a dataset of 10 million synthetic data, the largest in its industry.