By Michael Scott
The social behemoth’s most recent tools – Emu Video and Emu Edit- can evoke striking visuals from text, creating a generative artificial intelligence faceoff.
On Thursday, Meta revealed a preview of Emu Vido and Emu Edit, its two latest artificial intelligence (AI) tools, offering the initial accurate depiction of technology announced during September’s Meta Connect. Emu Edit is a tool that introduces inpainting, a unique approach to editing images. On the other hand, Emu Video allows users to make videos purely from text prompts.
The unveiling of the two AI tools underscores Meta’s strategic move. According to the technology conglomerate, this move still adheres to its more comprehensive vision for the Metaverse. The tools provide unique creative capabilities to attract many users, from professional content makers to people seeking novel means to communicate ideas.
Mainly, Emu Video shows the firm’s dedication to promoting artificial intelligence-enhanced content creation. Besides, it might be a significant competitor against famous names such as Pika Labs and Runway, who remain dominant in the space.
Emu Video Facilitating Text-to-Video Making
Emu Video embraces a two-step process for making videos using text prompts. It starts by creating an image based on the entered text and afterward creates a video acquired from the text and generated image. The strategy eases the video creation, evading the more intricate, multimodal strategies used to run earlier Make-A-Video tools.
Emu Video’s generated videos are restricted to a 512×512 pixel solution but indicate an outstanding consistency with the provided text prompts. Correctly changing text into visual narratives differentiates Emu Video from a majority of present models as well as commercial solutions.
The models are not available to the public. However, users can test a group of predetermined prompts, and the findings are pretty smooth, with few inconsistencies between frames.
Emu Edit Facilitates Image Editing with Inpainting
Together with Emu Video, Metal also demonstrated Emu Edit’s abilities. It is an artificial intelligence-enhanced tool developed to execute different image editing tasks based on artificial intelligence’s understanding of natural language commands.
Emu Edit permits people to edit images with significant flexibility and precision levels. Meta’s research paper for the tool mentions that Emu Edit is a multi-task image editing model that sets the latest results in instruction-founded image editing. This highlights its capability to implement intricate editing instructions correctly.
Diffusers, an improved artificial intelligence technology made famous by Stable Diffusion, improves Emu Edit’s accuracy. The approach ensures the maintenance of the original image’s visual integrity.
Meta’s objective of creating artificial tools such as Emu Edit and Emu Video symbolizes its plan to develop technologies critical to developing the Metaverse. This entails Meta AI’s development and the unveiling of multimodality in augmented reality devices.
Editorial credit: Ascannio / Shutterstock.com