Amazon Releases Second Version of its AI Image Generator With More Controls
Amazon has launched a second iteration of its in-house AI image generator for Amazon Web Services (AWS) customers.
Amazon’s image generator is called Titan and its second version is simply called Titan Image Generator v2.
The new model has the ability to “guide image creation using reference images, edit existing visuals, remove backgrounds, generate image variations, and securely customize the model to maintain brand style and subject consistency.”
The update means that as well as using a text prompt, users can provide an image that Titan will follow the layout and structure of.
Users can also precisely control color palettes by providing a hex code in the prompt and there is an automatic background removal tool.
The fine-tuning tool allows users to take one character or item — for example, a cartoon cat or product like a handbag — and then have that same item across multiple images.
“Titan Image Generator v2 can intelligently detect and segment multiple foreground objects,” says AWS principal developer advocate Channy Yun.
“With the Titan Image Generator v2, you can generate color-conditioned images based on a color palette. [And] you can use the image conditioning feature to shape your creations.”
Amazon has previously told PetaPixel that Titan won’t hurt photography saying that its image generator will supplement rather than replace the original content photographers and creative digital artists produce for their clients.
“The Titan Image Generator would really be meant to augment what photographers are always doing because, and based on my observations and what I’ve seen at the enterprise level, is most people want new and original content, especially when they’re looking at photographing new events or new people. They’re going to still want photographers,” said Sherry Marcus, Director of Applied Science in the Generative AI Services Organization at AWS in an exclusive interview with PetaPixel.
Tech Crunch notes that Amazon remains vague about the data it used to train its Titan models; saying only that it uses a mix of proprietary and licensed data.
AI companies have become very wary of revealing the data they use to train models with as training material runs dry and firms begin to guard their sources.
But it’s also to protect from lawsuits. If a tech company is found to be using copyrighted work then the IP owner can sue since the controversial issue of using copyrighted data to train AI models remains an open legal issue.
Image credits: Amazon
