Getty Images announced this week the launch of a new model for its AI services and generators. According to the platform, the update improves content generation, reaching around 6 seconds for 4 images, double the previous performance.
NVIDIA technology powers the updated model, delivering improvements in image quality and generation speed. Text instructions, or prompts, are more precise and can be longer. The definition quality of images upscaled to 4K resolution has also improved, and models can be customized to match each brand’s identity, style and content.
Building on the initial release of Generative AI by Getty Images and Generative AI by iStock, the update introduces features such as:
– Ultra-fast generation: The imaging speed now reaches about 6 seconds, doubling the performance of the previous model, which puts it at the forefront of the industry.
– Advanced definition in 4K generation: The generated images have better quality of detail and fidelity. In addition, they feature advanced resolution amplification (‘upscaling’) and improvements in the definition of 4K images.
– Better comprehension and agreement with more detailed prompts: Higher level of detail text prompts result in images that better match the descriptions provided in the text.
– Longer prompts: Text prompts can now be more complex and longer, up to 250 words.
– Advanced camera controls: For best results, you can now select the lens type and depth of field.
As part of the update, the company is also introducing AI-powered image modification capabilities, allowing users to tweak both AI-generated images in the tool and existing commercial stock photos in their libraries. These new capabilities allow users to add, change or remove elements in an image, as well as expand the dimensions or remove backgrounds with a single click. These capabilities are now available on iStock and will soon be available on Getty Images.
“We continue to push the boundaries of what’s possible with AI-generated imagery, and we do so responsibly,” said Grant Farhall, Chief Product Officer at Getty Images. “Our new model and the new capabilities we’re launching will give our customers even more opportunities to leverage AI. By combining this technology with our vast and incredible library of commercial photography and video, they’ll be able to bring their visions to life more efficiently and affordably than ever before, while ensuring commercial use safety and Getty Images’ distinctive legal protection.”
According to Getty Images, advances in this model benefit users across a range of industries, offering tools that streamline creative workflows and enable safe experimentation with generative text-to-image AI.
The company recently launched an option that allows its corporate clients to customize AI models, incorporating distinctive elements of their brands and producing images free of trademarks and copyright-protected logos.
The model was trained exclusively on Getty Images’ library of ‘creative’ images, which are suitable for commercial projects. The AI generator from Getty Images and iStock complements the company’s library of images and videos and its custom content offering, enabling customers to enhance their creative process and find the right visual content for their needs.
The model update leverages the NVIDIA Edify architecture, part of NVIDIA Picasso, an AI foundation for building and deploying generative AI models for visual content.
Follow ADNews on Instagram and LinkedIn.
#ADDSomethingNew #ADNewsOficial #ADGroupBR
