00:00

NeuroViz

Democratising Advanced Medical Visualisation
through Augmented Reality & DICOM Parsing

01 — Abstract
The Problem
& The Solution

Medical imaging—especially MRI—presents scan data as flat 2D grayscale slices, demanding years of expert training to interpret. This severely limits patient communication and restricts medical students without access to advanced imaging workstations.

NeuroVizAR transforms standard DICOM files into interactive 3D holographic models overlaid on the real world through a smartphone camera. Built with Unity 2022 LTS, ARFoundation, ARCore, and the fo-dicom library, it targets medical students, radiologists, and patients in Tier-2/Tier-3 Indian cities—offering a cost-effective alternative to INR 5–15 lakh DICOM workstations.

Augmented Reality MRI Visualisation DICOM Unity ARFoundation fo-dicom Volume Rendering Medical Imaging
🏥
1:834
India's Doctor-to-Patient Ratio
vs WHO-recommended 1:1000
💸
₹15K
Target device cost
vs ₹5–15 Lakh DICOM workstations
📱
ARCore
Android AR Backend
+ ARKit for iOS devices
02 — Problem Statement
What We're
Solving

The Core Challenge

Radiologists and clinicians work with 2D image stacks requiring significant expertise to mentally reconstruct as 3D anatomy. Patients, students, and non-specialist doctors are often unable to understand these images without guidance.

Conventional DICOM workstations costing INR 5–15 lakhs are out of reach for clinics and colleges in Tier-2/Tier-3 cities—leaving an enormous gap in accessible medical visualisation.

Literature / Market Gap

HoloLens-based AR systems (Gsaxner et al., 2023) exist but cost tens of thousands of dollars. No affordable smartphone-native DICOM viewer with full 3D AR overlay exists for emerging markets.

Innovation

NeuroVizAR delivers volumetric MRI rendering directly on ARCore/ARKit smartphones using Unity's GPU ray-marching shaders—no specialized hardware. First-of-its-kind for Tier-2 Indian healthcare contexts.

Data Source

Brain MRI DICOM files from OpenNeuro (openneuro.org), processed via 3D Slicer, then parsed by fo-dicom 5.1.2 within Unity for pixel intensity extraction.

03 — Platform Vision
Beyond Doctors —
A Full Ecosystem

NeuroVizAR is not a radiology tool. It is a medical spatial computing platform serving every stakeholder in healthcare & education — across AR, 3D, VR/WebXR, and DICOM slice viewing.

Working Prototype

A fully functional 3D MRI brain viewer with annotated zone labels, MRI slice navigation (Axial / Coronal / Sagittal), real-time Three.js WebGL rendering at 61 fps — running live on web. AR overlay and VR headset modes are in active development.

View Code & Source on GitHub
04 — System Design
Architecture &
Technology Stack
Unity
v6.3 LTS
Primary development engine, C# scripting, volume shader
ARFoundation
v5.x
Cross-platform AR framework for plane detection & anchoring
ARCore XR
v5.x
Android AR backend — ARKit 5.1.x for iOS
fo-dicom
v5.1.2
DICOM file parsing & pixel metadata extraction in C#
Visual Studio
2022 CE
C# scripting IDE and Unity debugger
OpenNeuro
Dataset
Open Brain MRI DICOM dataset for testing & training
3D Slicer
v5.4
Advanced open-source medical image processing and 3D visualization.
System Workflow Pipeline
NeuroVizAR full system workflow Complete pipeline from DICOM input through Unity and Needle Engine, branching to mobile AR on Android and iOS, WebXR browser, and VR headset outputs. DICOM Input .nii.gz · .dcm series Module 1 — DICOM Parser fo-dicom · 16-bit to float Module 2 — Volume Builder Texture3D · Volumetric shader engine routing Unity 6.3 LTS ARFoundation AR · VR pipelines Needle Engine Three.js · WebXR API Cross-platform viewer Module 3 — AR Session AR Plane tracking Module 4 — AR Placement Spatial anchor Module 5 — Interaction Touch interaction 2D Slice Viewer 3-axis projection 3D Brain Viewer Annotated 3D model WebXR Scene Zero-install web session platform output ▲ ARCore Android AR Google ARCore Android Devices ◆ ARKit iOS AR Apple ARKit iOS Devices ⬡ Needle WebXR Browser Any WebGL Browser Device agnostic ⬡ WebXR VR Headset Immersive VR Room-scale End Users Students · Doctors · Researchers TECH STACK fo-dicom · Unity C# ARFoundation · ARCore · ARKit Needle Engine · Three.js · WebXR Output: Android AR · iOS AR VR Headset (Meta Quest) WebXR browser · shareable link
DicomLoader.cs — Core DICOM Loading Logic
using FellowOakDicom;
using FellowOakDicom.Imaging;

public class DicomLoader : MonoBehaviour {

  public List<DicomSlice> LoadSeries(string folderPath) {
    var slices = new List<DicomSlice>();

    foreach (var file in Directory.GetFiles(folderPath, "*.dcm")) {
      var dcm = DicomFile.Open(file);
      var pixels = new DicomImage(dcm.Dataset)
                     .RenderImage().AsClrBitmap();
      slices.Add(new DicomSlice(pixels));
    }
    return slices;
  }
}
05 — Results & Testing
What's Working
Right Now
DICOM Files Parsed 50+
Pixel Normalisation Accuracy 70%
Modules Done 3 / 6
WebXR Frame Rate Target 60 FPS
05 — Roadmap
Development Roadmap &
Strategic Vision
Project Inception Current: Implementation Phase Future Horizon
55% work complete
CORE FOUNDATION [DONE]
DICOM Data Pipeline & Environment
Successful integration of fo-dicom framework, development of the asynchronous DICOM loading engine, and 16-bit to 32-bit pixel normalization for precision medical rendering.
VISUALIZATION ENGINE [DONE]
3D Rendering & Multi-Plane Interactivity
Developed the real-time 61 FPS WebGL viewer with annotated brain zone labels and a precise MRI slice navigation system (Axial, Coronal, Sagittal).
AR INTEGRATION [IN PROGRESS]
Holographic AR & Volume Reconstruction
Implementing HLSL GPU-based volume ray-marching shaders and cross-platform ARFoundation tracking for ARCore (Android) and ARKit (iOS).
BEYOND 2026: THE FUTURE HORIZON [UPCOMING YEARS]
Advanced Clinical & Surgical Integration
Expanding the system to support real-time AI tissue segmentation, multi-user collaborative surgical planning, and direct cloud integration with hospital PACS systems.
06 — Academic Credits
The
Team
Project Guide
Dr. Aditya Sinha
Dept. of CSE · MUJ
Developer · PBL-2
Souvik Samanta
Reg. No. 2427030613 · IV Semester