Along with providing access to Gemini Pro to developers and commercial customers. Google also announced generative AI as their latest image Imagen-2. An improvement of Imagen-2 over Imagen is that it can now understand text better. If the prom asks for some text to be inserted into the image it wants to generate, it can be done precisely unlike in the previous model.
More interesting here is that Imagen 2 can write text in six more languages namely Chinese, Hindi, Japanese, Korean, Portuguese and Spanish and is not limited to English only. More language support will be done next year.
Each generated image also has a SynthID watermark that makes it easy to trace its origin. With this people can see meta data that will tell them that the image they are seeing is generated by AI.
With more advanced models, Google says higher-quality, more realistic-looking images can be generated. The ability to generate logos for businesses is also provided along with the ability to generate mockup images for the logo to be used on products.
Our previous experience using Imagen has to be admittedly a bit disappointing as the image quality generated is not as good as DALLE-3 or Midjourney. Like Imagen, Imagen-2 is accessible to Google Cloud customers through the Trusted Tester Program today.