ABSTRACT
The proliferation of counterfeit and substandard drugs has dominated the pharmaceutical industry in Nigeria. This menace has been fueled by the prevailing establishment of unregistered off-the-counter drug stores also known locally as “chemists”, ineffective detection methods at sea ports, airports, land borders, markets, and local vendors, lack of continuous sensitization of the dangers of counterfeit drugs as well as the insufficient commonization of detecting methods amongst the local populace. Prominent amongst the issues outlined, is the commonization of detection methods for counterfeit drugs. This project aims to develop a more effective detection method using visual analysis of drugs by utilizing qualitative features found on these drugs. Various platforms and software development environments were used in development of this project, and they include: Visual Studio Code, Kaggle, Streamlit.
The project takes shape with the collection of datasets of the drug sample for the computer vision model training. The datasets are divided into train and test datasets with the train dataset used to train the image detection/classification model. The model is tested for accuracy using the test dataset and fine-tuned to improve the error metric used. Optical technologies precisely- QR code is employed to augment the functionality of the computer vision model. These technologies are deployed to a web application that can be accessed by anyone with a smartphone and an internet access.
The image of a drug sample is taken by a camera and passed through the model which classifies the drug as authentic or counterfeit with a benchmark of accuracy. The result of the classifier proved that a computer vision model can indeed be used to verify the authenticity of a drug going into the Nigerian market or already in the Nigerian market. The QR code scanner is used to scan the QR code on the pack of the drug to verify the manufacturer’s identity. The results showed that while optical and track and trace technologies are impressive methods to combating counterfeit drugs, the application of effective computer models is a boost to effectively addressing the menace of counterfeit drugs.a
