"Create Stunning Four-Second Animations with Google’s Imagen 2"

Google has unveiled important updates to its family of image generation models, including Imagen 2, which now includes text-to-live capabilities that allow users to convert text prompts into animated images. The model also introduces image editing features—such as inpainting, outpainting, and digital watermarking—now available for general use.

Announced at the Google Cloud Next conference, Imagen 2's text-to-live functionality generates animated GIFs initially at 24 frames per second, with a resolution of 360x640 pixels and a duration of four seconds. Google has indicated plans for continuous enhancements to this feature.

During a press briefing, Google Cloud CEO Thomas Kurian explained, “Instead of having a static image of an object, like a car, users can now see a short animation of a moving vehicle. Organizations, particularly in media and advertising, are adopting this technology to boost user engagement.”

Imagen 2 is designed to create images with various camera angles and motions, while ensuring consistency throughout the animation sequence. It also incorporates safety filters and digital watermarks, addressing key concerns related to generative AI.

The newly public image editing features enable users to add or remove elements from photos, akin to Adobe Photoshop's generative fill or content-aware tools. Users can also expand the image borders for a broader view.

These updates are part of Google’s announcements regarding Vertex AI, its fully managed cloud AI platform. Launched in 2023, Imagen 2 is a product of Google DeepMind, created to generate photorealistic, high-resolution images from natural language prompts. It competes with other leading models such as OpenAI’s DALL-E, Midjourney, and Adobe Firefly, specifically aimed at helping enterprises produce images that align with brand guidelines and governance standards.