Case Study: IDNEO achieves >90% accuracy and halves development time for MDmulticard blood‑type reader with MathWorks

IDNEO Develops Embedded Computer Vision and Machine Learning Algorithms for Interpreting Blood Type Results

IDNEO worked with Grifols to automate interpretation of the MDmulticard, a rapid blood antigen test, but faced variability in band appearance, limited initial samples, and a tight schedule to deliver a prototype for clinical feedback. To meet these challenges they used MathWorks tools—primarily MATLAB with Image Processing Toolbox, Embedded Coder, MATLAB Compiler, and Statistics and Machine Learning Toolbox—to support rapid algorithm development and deployment.

Using MathWorks software, IDNEO developed image-processing, computer-vision, and machine-learning algorithms in MATLAB, generated production C code with Embedded Coder, and integrated the code into an Android app, with standalone test apps produced via MATLAB Compiler and CI testing through Jenkins. The MathWorks-based workflow produced a preproduction prototype now in hospital usability testing, exceeded the >90% accuracy requirement (zero false positives/negatives on the sample set), and cut project time from 24 months to 12 months while delivering an optimized system.

IDNEO

Marc Blanch

R&D Hardware Supervisor