AI-Powered
Raman Spectroscopy
Raman Spectroscopy
for Counterfeit Antimalarial Detection

Why Does This Matter?

Global Health Impact

Counterfeit medicines are a silent public health threat.

Substandard and counterfeit antimalarial drugs contribute to treatment failure, rising drug resistance, and millions of preventable illnesses worldwide. This project enables rapid, non-destructive chemical verification at the point of care, helping protect vulnerable populations where laboratory testing is unavailable.

Accessibility & Cost

High cost limits impact.

Commercial Raman systems are expensive and infrastructure-heavy. This work shows that affordable hardware and open-source tools can deliver meaningful chemical analysis outside traditional laboratories.

Rapid Field Deployment

Real-world use demands speed and simplicity.

The compact, portable design allows fast drug screening in clinics, pharmacies, and field settings without sample preparation or specialized training.

Intelligent Analysis

Automated insight reduces human error.

Machine learning transforms raw Raman spectra into reliable classifications, enabling consistent drug authentication without expert interpretation. This approach improves accuracy, scalability, and usability, making advanced chemical analysis accessible to non-specialists in high-impact environments.

How Does it Work?

Optical Excitation

Laser Excites Molecular Vibrations

A laser illuminates the sample, causing most light to scatter elastically and a small fraction to scatter with shifted energy, producing Raman signals used for chemical identification.

Collection of Raman Signal

Light Is Collected and Separated

Backscattered light is gathered by lenses and directed through a diffraction grating, which separates wavelengths so that the detector can record intensity at each Raman shift for analysis.

Detection & Digitization

CCD Captures and Converts Signal

The CCD captures dispersed Raman light as analog voltages. The microcontroller uses its ADC to convert these analog signals into digital spectral data for further processing.

Machine Learning Classification

CNN Classifies Raman Spectrum

The Raspberry Pi runs a trained CNN model on the digitized spectrum, classifying whether the sample is legitimate or counterfeit by comparing learned spectral features in real time.

Loading…