As attendees arrived at the SpringML and Google Cloud booths, we did the typical meet and greet about who we are and what we do – but quickly took it to the next level by asking them, “Pick a theme or topic.” We then brought their topic to life with AI models.
How it works
Artificial intelligence models generate photo-realistic images by extracting meaning from textual data and using the knowledge from real images to build up textures, tones, and desired concepts. For example, when presented with the phrase “a close-up of a yellow flower,” the model will render an image including a petal closeup with the appropriate background blurred out. However, if given another description, such as “a man riding a rollercoaster,” it will generate an image of a male subject on what looks like an amusement park ride with individuals in the background cheering him on. The results are always incredibly realistic and accurate down to small details specified in the text, such as “wearing blue jeans.”
As AI advances, with technology like Google Imagen, so do the limits of our creativity and imagination, with the ability to create detailed photos from textual data inputs.
See what was created
a giraffe made out of asparagus skiing down a hill of mash potatoes
Someone tying a tornado in a knot, digital art
Someone tying a tornado in a knot, digital art
a burrito monster destroying new york city
Astronaut drinking coffee on the moon watching the earth, digital art
a cat submarine chimera, digital art
halloween jack o’lanterns eating pumpkin pies in a pumpkin patch
Coffee-creature eating a Asiago Bagel while getting hit by a UTA bus
What’s next
This technology is still in its early stages but is rapidly advancing. In five years, we may see project managers (like myself) use this technology to create entire proposals composed only of text which then get turned into corresponding visual images that are more aesthetically pleasing than straight text could ever be. With the new Google Imagen tool, we are excited to continue to play around with features such as these to see what results we get when we input various phrases or descriptions.
Are you ready to see it in action? Or see what is next in technology advancements? Meet us at the Google Government Summit on November 15th in Washington, DC, as we will have another engagement you won’t want to miss on Vision AI.
