2023 – Present · product

Slice-AR

Real-time 3D/AR viewer for CT and MR volumetric datasets. Runs entirely on-device with no cloud or network access required.

Unity Swift AR DICOM Metal Three.js Node.js WebSocket Medical Imaging GPU Rendering

The Problem

CT and MR scans produce rich volumetric data that traditional 2D viewers flatten into slices. Clinicians scroll through hundreds of images to build a mental model of a structure that is inherently three-dimensional. Industrial inspectors face the same constraint with NDT scan data.

Existing 3D viewers require specialist hardware, complex installation, or cloud upload of sensitive patient data. None of them exploit the spatial tracking capabilities already built into modern mobile devices.


The Approach

Slice-AR treats the device itself as the cross-section plane. Hold an iPad up to a scan, and the display shows exactly the slice at that physical orientation, updating in real time as you move.

The system runs entirely on-device: data is loaded from local storage, processed into a volume bundle, and rendered with GPU ray marching. No cloud. No network. No specialist hardware beyond a standard tablet.

Slice-AR iPad app: real-time cross-section slicing

Volume rendering uses a custom cross-section shader with Hounsfield Unit (HU) windowing, so tissue types and densities can be isolated with a slider rather than requiring radiologist presets.


Display Modes

iPad: Device as Cross-Section Plane

The iPad uses its built-in position tracking to act as a physical cross-section plane. Move the device through the volume and the slice updates instantly.

The app is publicly available on the App Store (multi-lingual: EN, DE, FR, ES, IT, JP):

Download on the App Store

Key characteristics:

  • On-device import and processing; no cloud or network access
  • Built-in ARKit position tracking drives the slice plane
  • Annotation tools for marking regions within the volume
  • Runs on standard iPad hardware

Stereo 3D: Holographic Display with Tracked Frame

A second display mode uses an Illumetry IO stereo 3D display, which renders the volume as if occupying the same physical space as the viewer. An Antilatency Alt Tracker turns a physical frame into a tracked input device, allowing the user to reach into the volume and define the cross-section plane spatially.

Slice-AR: stereo 3D display with Antilatency tracked frame
Slice-AR: tracked frame and stylus acting as an x-ray torch

This mode is aimed at clinical and research environments where a shared, hands-free view of the volume is useful.


System Architecture

Data flows through five stages:

DICOM / raw image series
Local import and processing
GPU volume rendering
Spatial tracking
Display output

Volume data never leaves the device. Only slice plane coordinates travel over the local WebSocket relay when connecting to observer displays.


Audience

The tool is useful across several contexts:

  • Healthcare professionals who benefit from interpreting scan data spatially: surgeons, specialists, and radiographers
  • Patients who need to view their own scan data without installing complex viewers
  • Educators and learners building spatial understanding of anatomy
  • Industrial NDT (Non-Destructive Testing) professionals working with CT scan data

Status

The iPad app is published on the App Store under Roy Rodenhaeuser. The stereo 3D and observer display components are in active development as part of the broader Slice-AR ecosystem, which includes a web-based observer interface and a native iOS rebuild.