Stunad Spatial

Stunad Spatial explores point clouds as a primary representation for spatial computing. Instead of treating them as an intermediate step toward meshes, the system focuses on using point clouds directly as navigable spatial environments.

These environments can be explored in full six degrees of freedom, allowing users to move through reconstructed imagery or scanned environments and perceive spatial depth that normally remains collapsed inside flat media.

The system supports multiple sources of spatial data that can produce point cloud representations.

• Monocular image → depth reconstruction
• Photogrammetry scans
• Public spatial datasets
• Captured XR camera frames

Image reconstruction is one experiment within this system, where a single photograph can be transformed into a spatial representation that can be explored in XR.

Depth estimation runs locally on Meta Quest hardware using ONNX inference through Unity Sentis. The depth map is reprojected into 3D space using camera intrinsics, transferred into GPU buffers, and rendered as a volumetric point cloud.

The project focuses not only on reconstruction but on exploration. The resulting point cloud becomes a navigable environment that can be explored in full 6DOF.

Custom rendering systems manage large point sets and stream spatial chunks in order to maintain performance on standalone XR hardware.

Large architectural scans were used as experimental testbeds. The Pantheon point cloud dataset, sourced online, was used to explore how historical architecture could be experienced as a navigable point cloud environment.

These experiments investigate alternative XR workflows where environments emerge from captured spatial data rather than manually modeled geometry.

• Reliving memories through spatial photographs
• Exploring architectural scans
• Experimental spatial storytelling
• Research into point clouds as an XR medium

Point clouds are becoming a fundamental representation for spatial computing, digital twins, and XR capture pipelines. By treating them as a first-class medium rather than a preprocessing step, Stunad Spatial explores how spatial data can become the foundation for immersive experiences.

Photography captures moments. Spatial computing might allow us to step inside them.