Google’s AI text-to-image generator Imagen is gaining a new feature that allows it to generate short four-second-long videos from animated images, similar to GIFs.
Imagen 2 is an enhanced version of the tool, which Google announced at its Cloud Next conference in Las Vegas (via TechCrunch).
Google calls the new feature “text-to-live image,” and it can create four-second-long videos from user prompts. Examples shared by Google include prompts like “Aerial view of a mountain range,” “Capture a pot of Thai curry as it comes to a boil,” “Rocky coast, yellow then firm, crashing waves,” and more.
Google says that the tool will create live images at 24 frames per second (fps) and at 360 x 640-pixel resolution, with improvements to come in the future. All generated live images will have invisible SynthID tags to alert others about the image’s AI birth, paired with filters for safety.
It’s unclear why Google is working on an AI tool that generates short, low-resolution GIFs when its competition, like OpenAI’s Sora, can turn text into hyperrealistic videos. Google’s own Lumiere is also capable of generating higher-quality videos than Imagen 2. According to the tech giant, the tool is primarily for businesses, and it wants to be an enterprise application, likely for generating quick shots for ads.
Imagen 2’s text-to-live image capabilities should be generally available in the future.
Apple is also reportedly working on a similar tool called Keyframer that can generate animations from a single image by providing text instructions. You can read more about it here.
Image credit: Shutterstock
Source: Google Via: TechCrunch
MobileSyrup may earn a commission from purchases made via our links, which helps fund the journalism we provide free on our website. These links do not influence our editorial content. Support us here.
