If you’ve ever bought a sofa from an online store, have you ever thought about the house you see in the background? When launching a new collection, furniture brands usually spend a lot of money on photo shoots, a tedious and expensive process because it’s not easy to move furniture around.
That’s why French startup Presti, which uses generative AI to turn single product images into lifelike photographs, has already raised $3.5 million in a seed round led by Partech and with participation from several angel investors.
“We immediately got on the phone and spoke to 50 potential users,” Presti co-founder and CEO Nabil Toumi told TechCrunch, “and they all said the same thing: creating product visuals was a very time-consuming, expensive process, and there was no easy solution to create these photos. At the same time, it was really the most important asset for brands to create a unique identity and sell online.”
Presti translates this to:
to this:
Initially, Presti didn’t focus on furniture companies. But the startup quickly realized that furniture companies face a particularly tough problem.
“For the photo shoots, nice houses had to be rented, the goods had to be transported, so there were high logistical costs associated with them. And those photo shoots were planned months in advance, ultimately costing hundreds of thousands, sometimes millions of euros a year,” Tumi said.
Presti uses Stable Diffusion XL as its core foundation model, which has been retrained and tuned to be specifically suited for product imagery in the furniture industry.
Initially, the team tried to use the standard version of Stable Diffusion XL, but they quickly realized there were problems. “If you add legs to the sofa, the back will become distorted,” says Toumi. Similarly, getting the perspective right was difficult; for example, the wall behind the sofa needs to be parallel to the sofa.
“At the same time, what’s really important is the dataset that we trained the model on. Right now, in our industry, we have over 75,000 super-high-quality photos of furniture. We can use this to train the model and enhance the learning process for the specific use case of this type of photography,” Toumi says.
Presti didn’t want to stop at generating backgrounds: customers can also add accessories to their products. For example, if you generate a product photo for a new sofa, you can add cushions. These cushions cast realistic shadows on the sofa, so it doesn’t look like it was added in Photoshop.
Similarly, furniture brands typically have multiple variations of the same model, with different textures and colors. It’s still a work in progress, but Presti hopes that customers will be able to use its tools to swap out materials. That will make new product launches much easier for companies that use Presti. Freelance photographers won’t be thrilled with the new offerings, though.
Presti works primarily with mid-sized furniture companies, but also has a strategic partnership with Maisons du Monde, one of France’s largest furniture retailers. In addition to Partech, investors in the startup include Thibaud Elzière, Maxime Brousse, Abou Laraki, Rémi Lemonnier and Julien Hirth.
