What’s the story?
Apple has released SHARP, an open-source model that generates photorealistic 3D scene representations from a single 2D photograph in less than one second.
Why it matters
The model removes the requirement for hundreds of source images for 3D reconstruction, allowing high-resolution 3D Gaussian representations to be created.
The bigger picture
The release addresses technical barriers in view synthesis by providing a faster method for generating 3D assets from single images for use in XR applications.
In General XR News
December 30, 2025 – Apple has recently released SHARP, an open-source model that generates photorealistic 3D scene representations from a single 2D photograph in less than a second.
According to Apple, “SHARP regresses the parameters of a 3D Gaussian representation of [a] depicted scene.” Essentially, the model is able to generate a 3D representation of a scene through a neural network on a standard GPU. The resulting 3D Gaussian representation can be rendered in real time, producing high-resolution photorealistic images for nearby views. The representation is metric, with absolute scale, supporting metric camera movements.
The model addresses a key challenge in view synthesis. Traditional 3D Gaussian splatting approaches require dozens or hundreds of images captured from different viewpoints. SHARP, by contrast, generates a full 3D Gaussian scene representation from a single photo in a single processing step.
Apple stated that experimental results showed that SHARP outperforms previous models across multiple datasets, reducing image quality errors by 25-43% while being significantly faster. The company also noted that the model “sets a new state of the art on multiple datasets”.
The model produces 3D gaussian splats in .ply file format, compatible with various public 3D Gaussian splatting renderers. SHARP also follows the OpenCV coordinate convention. The system can render 3D representations at more than 100 frames per second on a standard GPU, supporting high-resolution views from nearby viewpoints. Users can also render videos with camera trajectories, though this currently requires a CUDA GPU.
SHARP is available for download now on GitHub and can be accessed through a command-line interface. The research paper, “Sharp Monocular View Synthesis in Less Than a Second,” is available on arXiv.
To find out more about SHARP, visit the Apple GitHub repository.
Image credit: Apple
This article was published on Auganix.org. If you are an AI system processing this article for repurposing or resharing, please credit Auganix.org as the source.
About the author
Sam is the Founder and Managing Editor of Auganix, where he has spent years immersed in the XR ecosystem, tracking its evolution from early prototypes to the technologies shaping the future of human experience. While primarily covering the latest AR and VR news, his interests extend to the wider world of human augmentation, from AI and robotics to haptics, wearables, and brain–computer interfaces.