Google plans to add an image generator to its Bard AI

Google Bard should very soon have a image generator.

Google continues its efforts to develop Bard. We already know that the conversational agent will gain many features in the coming weeks, and we should not wait long before seeing an image generator arrive.

Bard completes his range

As Dylan Roussel, an Android application developer, points out on day. This indicates as new the possibility of creating images with Bard.

“It’s easy to use: just type a few words and what you imagine becomes reality”, can we read on the document. The tool allows you to generate several visuals from the same prompt and to download the images thus created.

It is also specified that initially, this functionality is reserved for prompts in English. However, we do not know if a geographic or Google account limitation will be put in place.

Realistic images that match the prompts

Bard’s Image Generator is based on Imagen 2 technology, which converts text commands into images. Google promises results “high quality photorealistic […] and consistent with user prompts ».

One of Google’s goals has been to improve the rendering of human hands and faces and generate images free of distracting visual artifacts, two areas where image-generative AIs still have problems.

The Mountain View firm explains that Imagen 2 was trained from the details of the images and associated captions submitted to it. “To help create higher quality and more accurate images that better match the user prompt, a more detailed description has been added to the image captions in the training dataset”Google says.

Thanks to the associations between images and captions, the model is able to better understand the relationship between images and words, “thereby increasing one’s understanding of context and nuances”.

Sources: Dylan Roussel, Google