Case Study: Longwood University achieves rapid, low-cost diabetes risk prediction with MathWorks (MATLAB and Wavelet Toolbox)

Longwood University Researchers Use Wavelet Analysis to Assess Diabetes Risk

Longwood University researchers set out to predict early type 2 diabetes risk by quantifying fluctuating asymmetry in paired fingerprints, but faced limited time and budget to test multiple analysis techniques. To accelerate development they used MathWorks’ MATLAB and Wavelet Toolbox to experiment with wavelet-based approaches instead of building analysis tools from scratch.

Using MathWorks’ MATLAB and Wavelet Toolbox, Dr. Bjoern Ludwar cropped 340 fingerprint scans (257 diabetic, 83 control), tested wavelet decompositions, selected a Haar transform to produce 36-element feature vectors, and computed Euclidean distances between opposing fingers (five values per patient). The MathWorks tools let the team get initial results in days rather than months; the pilot study found a significant correlation between fingerprint FA scores and diabetes risk, identified the most predictive fingers, and suggested a lower‑cost alternative to genetic testing while enabling concise, reproducible algorithms and plans for a larger study and a future mobile app.

Longwood University

Bjoern Ludwar

Assistant Professor