Qu-Alz

An experimental research system combining brain MRI processing, stage classification, and quantum-derived entropy features.

Qu-Alz began with a research question: could MRI-based stage classification be paired with a quantum-derived signal to support research into earlier Alzheimer's risk assessment?

The resulting prototype connects medical-image processing, machine learning, quantum computation, and a web interface for reviewing the outputs together. It was first developed for SEA Quantathon 2025, where it placed first runner-up. The team later developed the work into a conference paper accepted for presentation at ASEAN IVO Forum 2025.

Research workflow

Brain MRI
    -> preprocessing and region segmentation
    -> four-stage impairment classification
    -> MRI feature encoding
    -> reduced density matrices
    -> von Neumann entropy
    -> Entanglement Entropy Score
    -> combined research view

The system includes:

• PyTorch-based MRI preprocessing and segmentation;
• classification across no, very mild, mild, and moderate impairment classes;
• a quantum path using a ZZFeatureMap, density matrices, and von Neumann entropy;
• an Entanglement Entropy Score considered alongside the image-model outputs;
• a Next.js interface for uploading an MRI image and reviewing the generated results;
• notebooks, scripts, trained prototype artifacts, tests, and generated result plots.

Presented results

The competition presentation framed the system as a combined pipeline rather than a standalone model. The repository retains the outputs used to discuss that pipeline:

• class-level EES comparison;
• EES visualization;
• stage-classification confusion matrix;
• sample segmentation and classification outputs.

These are experimental prototype results. The project did not conduct clinical evaluation or establish diagnostic performance.

Run locally

Research pipeline

cd quantum-risk
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
python test_quantum_ees.py

Additional usage notes are available in quantum-risk/PIPELINE_USAGE.md.

Interface

cd interface
npm install --legacy-peer-deps
npm run dev

The interface is a review surface for the prototype outputs; it is not a clinical application.

Recognition and presentation

• First runner-up, SEA Quantathon 2025
• Conference paper accepted for presentation at ASEAN IVO Forum 2025: Qu-Alz: Quantum-Enhanced Early Prediction of Alzheimer's in Southeast Asia
• Project demonstration
• CERN Open Quantum Institute event report
• Official ASEAN IVO Forum program
• Presentation archive

Team

Qu-Alz was developed by Roi Victor Roberto, Sayed Tahlil Hossain, Dimas Sakti Widyatmaja, and Mutawally Nawwar.

License

Licensed under the MIT License.