A startup called Fyusion Inc. is pitching its three-dimensional imaging software that uses machine learning to enable anyone to create extremely realistic 3-D images of people and objects from a few smartphone camera snaps.

3-D images are popularly used in digital marketing, but Fyusion said use of the technology isn’t very widespread because it has always been very expensive to create.

Traditional 3-D imaging methods, for example, involve using costly laser-scanning tools to create accurate representations of products. Alternatively, it’s possible to use a method known as “photo-stitching” to create 3-D images, but that often results in poor-quality images, according to the company.

What Fyusion has done is apply machine learning to the photo-stitching process in order to make its 3-D representations far more vivid and accurate.

“First, Fyusion uses a deep network to promote each source view to a layered representation of the scene, advancing recent work on the multiplane image representation,” the company explained. “Fyusion then synthesizes novel views by blending renderings from adjacent layered representations. The result is a 4,000x decrease in the number of images needed to produce a 3-D image, making it easy for anyone to create high-quality 3-D images using only a smartphone.”

The company’s software is targeted at retailers, allowing them to show off a “vivid and realistic” representation of their products, so customers feel as if they’re experiencing it before they buy.

Fyusion is demonstrating its software at the annual ACM SIGGRAPH computer graphics technology conference in Los Angeles next week. The software is available as a consumer app for both Android and iOS. In addition, the company offers an enterprise-grade version of its software with more features for marketers in industries such as automotive and fashion.

“The most compelling virtual experiences completely immerse the viewer in a scene, and a hallmark of such experiences is the ability to view the scene from a close interactive distance,” Pantelis Kalogiros, Fyusion’s co-founder and vice president, said in a statement. “This is currently possible with synthetically rendered scenes, but a high level of image ‘intimacy’ has been very difficult to achieve for virtual experiences of real-world scenes. It is critical to maintain the accuracy and photographic fidelity of the digitized 3-D scene as the camera view point is changed. Our new technology is the best in the world at this.”

Image: Fyusion