It’s not hard to imagine your picture-perfect 2023 Lexus RX or 2023 Lexus RZ 450e luxury utility vehicle. But, it’s even better to envision it on your own terms – especially amid the hustle and bustle of the New York International Auto Show (NYIAS).
Through the power of artificial intelligence, guests attending NYIAS at the Jacob Javits Center from April 7-16 are able to type in a prompt for the Lexus RX or RZ, placing either vehicle in a setting limited only by their imagination, and manifest that vision into existence.
“Lexus is focused on creating innovative ways to connect with our guests, and this time, we’re using Generative AI technology at the New York International Auto Show to engage with them in a unique and very personalized way,” said Vinay Shahani, vice president of Lexus Marketing. “As we’ve shown with the recent launches of the all-new Lexus RX and RZ, intuitive technology is a key pillar of the brand, and this experience featuring these two vehicles provides guests the opportunity to explore what it would look like to have the RX and RZ as a part of their everyday lives and their dreams.”
Also Read: Trimble Launches First Industry Dwell Time Metrics for Fleet Management
At the Lexus display, guests are guided through prompts that generate images that appear on a 98-inch screen, featuring either the Lexus RX or RZ. Afterward, guests can send themselves their images via email, allowing them to bring their art home with them with nearly any backdrop they could fathom – even in a galaxy far, far away.
To develop the AI art tool, Lexus collaborated with data scientists and machine learning engineers at Toyota Connected North America (TCNA), a data, cloud, artificial intelligence, machine learning and software center of excellence headquartered in Plano, Texas. TCNA leveraged state-of-the-art Generative AI models such as Stable Diffusion and ControlNet architectures to train a proprietary Generative AI model to generate photorealistic or artistic style images – all by way of text input. The model was trained on more than 500 photos of each vehicle, capturing every detail imaginable.
Additionally, the stable diffusion model was trained on LAION-5B dataset from non-profit organization LAION to use for backgrounds, artistic styles and detailing.
SOURCE: PR Newswire