3D can be an extremely valuable asset to brands and creatives, providing immersive, engaging experiences while streamlining the design process.
However, it can be costly, time-consuming, and difficult to implement efficiently – rendering it impractical for daily business needs.
Nvidia is once again making waves in generative AI; and is looking to establish itself in this emerging realm through GTC 2024 with their Nvidia Edify multimodal generative AI model capable of producing 3D content and in partnership with Shutterstock and Getty Images on Edify-powered tools.
Shutterstock has announced early access to an Edify API that creates 3D objects from text prompts and images to build virtual scenes in virtual reality environments.
Getty is also expanding their gen AI service with fine-tuning options so enterprise customers can create visuals according to brand guidelines and aesthetic styles.
Developers will soon be able to test these models through Nvidia NIM, an inference microservices suite introduced at GTC.
“3D Asset Generation is among the latest capabilities Edify offers developers and visual content providers, giving them more creative control over AI image generation,” wrote Gerardo Delgado, director of product management at Nvidia in a blog post about this latest capability.
Getty fine-tuning Edify for specific brands (Sam’s Club, Mucinex and Coca-Cola).
One of the key challenges associated with generation AI is finer control over image outputs from AI algorithms.
Getty announced at CES 2019 Edify-powered APIs for inpainting and outpainting to help address this problem, providing features such as inpainting (which allows users to add objects) and outpainting (which expands canvas size). Both these features can now be found on Getty’s website and iStock.com.
Starting May, Edify will offer new services enabling companies to tailor it specifically to their brand and style. This no-code, self-service method enables brands to upload proprietary datasets, review auto-tags, submit fine-tuning parameters and review results prior to deployment.
Developers will soon gain access to Sketch, Depth and Segmentation features that enable users to submit drawings that guide image generation; copy compositions from reference images using “depth mapping”, and segment images by adding, removing or editing characters and objects.
“Getty Images continues to expand the capabilities of its commercially safe gen AI service, which offers users indemnification against the content they produce,” writes Delgado.
Getty’s Gen AI tools are used by leading creatives and advertisers, according to the company, such as:
Dentsu Inc: Dentsu is using Nvidia Picasso to fine-tune Getty’s model for membership retail giant Sam’s Club, while supporting Manga Anime For All which generates anime-type content for marketing use cases.
McCann: McCann’s creative agency utilized gen AI to develop an interactive game for Mucinex cold remedies; users are invited to engage with its surly mascot Mr. Mucus through this feature.
WPP: WPP, the marketing and communications firm, has joined The Coca-Cola Company in utilizing Getty’s model to craft custom visuals for its iconic soda brand.
