Google Cloud has rolled out Imagen 2, an enhanced version of its text-to-image generation technology, available to Vertex AI customers on an allowlist basis. This latest iteration enables the creation and deployment of highly realistic images through user-friendly tools and a fully managed infrastructure.
Powered by technology from Google DeepMind, Imagen 2 boasts improved image quality and offers a suite of features designed for various specific applications.
Key attributes of Imagen 2 include:
– Advanced image generation: Imagen 2 excels in producing high-definition images from textual prompts, catering to a wide range of user needs.
– Multilingual text rendering: Addressing common obstacles, Imagen 2 supports precise text rendering in several languages.
– Logo creation capabilities: Businesses can use Imagen 2 to generate diverse, realistic logos and apply them to products, apparel, business cards, and more.
– Enhanced captioning and Q&A features: The model’s sophisticated image understanding enables it to generate descriptive captions and answer complex questions about the elements within an image.
– Expanded language support: Imagen 2 now includes preview support for six additional languages, with plans to add more in early 2024. This expansion also encompasses the ability to translate between the input prompt and the output.
– Robust safety measures: Adhering to Google’s Responsible AI principles, Imagen 2 integrates safety features, including safety filters and a digital watermarking service, to promote responsible usage.
Designed for Enterprise Applications
Imagen 2 on Vertex AI is tailored to meet enterprise-grade standards, ensuring reliability and governance similar to its earlier version. New functionalities such as enhanced image and text rendering, logo creation, and reinforced safety protocols position Imagen 2 as a comprehensive solution for creative image generation in business contexts.
Prominent companies like Snap, Shutterstock, and Canva have already incorporated Imagen into their creative workflows.
Chris Loy, Director of AI Services at Shutterstock, remarked on the importance of diversity in the creative process and the integration of cutting-edge technology into their image generator and editing features, provided it is built on responsibly sourced data.
Danny Wu, Head of AI at Canva, shared their ongoing efforts to leverage generative AI to revolutionize design and enhance creativity. With Imagen,
their over 170 million monthly users can take advantage of the improved image quality to enhance their content creation on a large scale.
As Imagen 2 begins to influence the creative sector, organizations are being encouraged to explore its capabilities. Google Cloud is optimistic that users will utilize the new features of Imagen 2 to boost their creative projects and build upon the achievements of the original Imagen.
This introduction of Imagen 2 signifies a significant advancement in the realm of AI-driven creative image generation, offering enterprises a powerful tool for transforming ideas into vivid, realistic images. The broad adoption of Imagen by leading companies in the creative industry underscores its potential and effectiveness in augmenting the creative process with AI technology.
