Apple created AI that turns a photo into a photorealistic 3D view in seconds and then open-sourced it
Apple has introduced SHARP, an AI model that rapidly converts single 2D photos into realistic 3D scenes. This open-source technology can transform AR and VR experiences while inviting broader developer collaboration.
Apple has quietly dropped one of its most impressive AI research projects to date, a model that can turn a 2D photo into a lifelike 3D scene in less than a second. The system, called SHARP, short for Sharp Monocular View Synthesis in Less than a Second, is detailed in a newly published research paper and made publicly available on GitHub.
In short, SHARP lets you take a simple image and view it from slightly different angles as if you were standing in the scene yourself. It doesn’t just show fake depth with filters, it actually creates a 3D representation of the photo that looks realistic when viewed from multiple perspectives. For Apple, it’s the latest demonstration of its growing expertise in advanced computer vision and AI research, areas that underpin everything from iPhone photography to the company’s work in augmented and virtual reality.
From flat photo to 3D view in under a second: how it works
So, how does SHARP accomplish this almost magical transformation? According to Apple’s research team, the model analyzes a single image and predicts what the scene might look like in three dimensions. It then creates a lightweight 3D map made up of millions of tiny data points called a Gaussian. Each acts like a blurry bubble of color and light located in Gaussian space. When rendered together, these points create a realistic, dynamic scene that can be viewed from surrounding angles, complete with smooth depth changes and parallax effects.
This means you can move your virtual camera slightly left or right and watch the background subtly change, just like when you tilt your head in real life. Unlike most traditional 3D reconstruction methods, which require dozens or even hundreds of photographs of the same subject taken from different angles, SHARP manages the entire process with just a single photograph, and completes it in under a second using a standard GPU.
Apple researchers say this one-pass system makes SHARP different from previous techniques that relied on slow, per-view optimizations. In other words, SHARP doesn’t require hours of processing or specialized hardware, it just works faster.
But there is a catch
SHARP is not intended to create a fully explorable 3D world or to fill in parts of an image that have never been captured before. It focuses on local viewpoint changes, meaning users can view a scene from slightly different angles but not completely rotate around it.
Apple acknowledged this limitation in its research paper, noting that the model is designed to maintain realism rather than add confusing new details. The goal is credible depth, not imaginary reconstruction.
Yet, even within these constraints, the possibilities are vast. SHARP can deliver new types of spatial photos, AR and VR experiences, or advanced visual effects that bring everyday images to life. Apple already offers similar features on devices like the iPhone 15 Pro, where users can capture “spatial photos” or apply depth effects to standard shots. SHARP can take those experiences to the next level, making 3D photography faster, more accurate, and more accessible.
Perhaps the most surprising part of this release is that Apple has made SHARP open-source. By publishing code and research online, the company is inviting developers and academics to experiment, adapt and build on top of its work, a rare move for Apple, which generally keeps its cutting-edge projects secret.