Google has announced the Imagen 2 AI model with text and logo generation capabilities
Google has introduced an updated version of its Imagen neural network for image generation. Imagen 2 has expanded capabilities for creating texts and logos based on text prompts.
The new model was developed with the involvement of specialists from DeepMind, a division of Google specializing in artificial intelligence technologies.
According to the engineers, Imagen 2 generates better quality images. In addition, the model can now create images with text, for example, for advertising.
Another innovation is the generation of logos and their overlay on existing photos. This puts Imagen 2 on par with competitive models from OpenAI and Amazon. At the same time, Imagen 2 supports text in several languages, including English, Spanish, Chinese and Hindi.
According to Google, new learning algorithms allow Imagen 2 to understand longer and more detailed queries in English and other languages. All generated images are labeled with SynthID technology to prevent misuse.
