Apple’s Open-Source SHARP AI Model Converts 2D Photos To 3D Scenes Instantly

Apple has unveiled SHARP (Sharp Monocular View Synthesis), an open-source AI model capable of converting a single 2D photo into a photorealistic 3D scene in under a second. Developed by Apple Research, SHARP uses a feedforward neural network to predict depth and geometry, generating millions of 3D Gaussian points for realistic rendering. Unlike traditional methods requiring multiple images, SHARP achieves speed and accuracy with minimal input, preserving real-world scale and depth. Trained on millions of synthetic and licensed photos, the model opens new possibilities for AR, VR, design, and creative industries, with code available on GitHub.