Detection of Counterfeit Medicines Using Hyperspectral Sensing.

Area of Science:

• Pharmaceutical analysis
• Spectroscopy
• Machine learning

Background:

• The global pharmaceutical market faces significant losses due to counterfeit medicines, estimated at $200 billion USD.
• Current detection methods require technical expertise, lab equipment, and time, limiting field applicability.

Purpose of the Study:

• To develop a simpler, faster, and field-deployable method for detecting counterfeit medicines.
• To utilize hyperspectral sensing technology for identifying adulterated pharmaceutical products.

Main Methods:

• Visible-near infrared (350nm-1050nm) hyperspectral imaging was employed to capture spectral signatures of 24 medicine tablets.
• Counterfeit conditions were simulated by adulterating tablet powders with varying levels of calcium carbonate.
• Machine learning, specifically a multilayer perceptron classifier, was used to analyze spectral data.

Main Results:

• Hyperspectral sensing successfully differentiated between genuine and adulterated medicine tablets.
• The developed method achieved over 90% classification accuracy in detecting counterfeit samples.
• Spectral signatures varied predictably with increasing levels of adulteration.

Conclusions:

• Portable hyperspectral sensing presents a viable solution for rapid, on-site detection of counterfeit medicines.
• Integration with a spectral database can enhance the accuracy and scalability of field-level counterfeit drug testing.
• This technology empowers non-experts to identify potentially harmful counterfeit drugs, improving public safety.