Meta introduces Emu Edit and Emu Video: AI tools for image and video editing

Meta Platforms unveils Emu Edit, an AI-powered image editor, and Emu Video, a text-to-video generation tool, both based on the Emu foundational model. Know all about them.

Meta Platforms has announced two new AI-powered tools. Know details. (REUTERS)

Seeing the boom of artificial intelligence (AI), Meta Platforms has been making efforts to incorporate this technology into its suite of products, and its latest offerings include an AI-powered image editor, as well as a video generator. These products follow the unveiling of the multi-persona AI chatbot called AI Personas which the company debuted at its Connect 2023 event last month. The company also announced Emu, its first-ever foundational model for image generation. The new AI tools announced by Meta are based on the same model. Here's all we know about it.

Meta AI tools

In a blog post, Meta announced two new AI tools - Emu Video, and Emu Edit, and both of them are based on the Emu foundational model.

1. Emu Edit - This tool is an AI-powered image editor that promises to streamline image manipulation tasks and introduce precision image editing. The company claims it will be able to “precisely alter images based on text inputs.” Through Emu Edit, users can carry out a variety of tasks such as local and global editing, removing and adding a background, colour and geometry transformations, detection and segmentation, and more. To train the AI model, Meta used 10 million synthesized samples, including an input image, a prompt, and the output result.

Meta said, “In our evaluations, Emu Edit demonstrates superior performance over current methods, producing new state-of-the-art results in both qualitative and quantitative evaluations for a range of image editing tasks.”

2. Emu Video - Emu Video offers text-to-video generation capabilities based on diffusion models. According to Meta, the tool offers multiple input options - text only, image only, and both text and image. Generating the video takes a split approach. First, the images are generated based on the text prompts. Then, Emu Video generates a video based on both the previously generated image as well as the text prompt. This tool uses two diffusion models to generate 512x512 four-second-long videos at 16 frames per second.

Meta says both tools could potentially allow users without any technical expertise to generate and edit photos and videos. “While certainly no replacement for professional artists and animators, Emu Video, Emu Edit, and new technologies like them could help people express themselves in new ways—from an art director ideating on a new concept or a creator livening up their latest reel to a best friend sharing a unique birthday greeting”, the company says.