Google has announced an enhanced version of Imagen 2, integrated into its Vertex AI developer platform, at the annual Cloud Next conference in Las Vegas. This update brings Google's state-of-the-art image generative AI capabilities to application developers, enabling them to build next-generation AI products that transform users' imagination into high-quality visual assets within seconds.
The latest version boasts improved language comprehension and photo-realism, earning Google a leadership position in the recent Forrester Computer Vision Wave. This advanced text-to-image technology is now generally available on Vertex AI, enabling businesses to create images that match their specific brand requirements.
During the announcement, Google introduced several exciting features. First, Text-to-Live Image, currently in preview, allows marketing and creative teams to generate animated images, product visualizations, ads, GIFs, and storyboards from simple text prompts. An example of this was showcased, depicting a daylily opening in the morning dew and a pot of delicious stew simmering on a stove, bringing static images to life.
Second, Google announced the general availability of Digital Watermarking for AI-generated images, powered by Google DeepMind's SynthID. This ensures the integrity and authenticity of generated images, addressing concerns about the modification and misuse of digital content.
Additionally, Imagen 2.0 introduces new editing modes that provide users with enhanced control and creativity. Users can easily remove unwanted elements from images, add new elements, and even expand frame to get a wider view.
In addition to image generation and editing, Imagen on Vertex AI offers visual captioning, providing text descriptions of images, and Visual Question Answering (VQA), which delivers answers to questions about an image. These features further enhance the platform's versatility and potential applications across various industries.