Conversational Artificial Intelligence

17Counterfeit Pharmaceutical Drug Identification

Sajidha S. A.^1*, Aakif Mairaj², Amit Kumar Tyagi³, A. Vijayalakshmi¹, Nisha V. M.¹, Siddharth Nair¹, C.K.M. Ganesan¹, Ram Gunasekaran¹ and Hitarth Menon¹

¹School of Computer Science and Engineering, Vellore Institute of Technology, Chennai, India

²School of Sciences, Indiana University Kokomo, Kokomo, Indiana, USA

³Department of Fashion Technology, National Institute of Fashion Technology, New Delhi, India

Abstract

Counterfeit drug selling is an extremely dangerous problem around the world, which has escalated in the recent past. It has caused millions of deaths and serious complications in the health of people who consume them. It is not possible for the common man to detect subtle differences in the packaging and the back-strip labels of drugs. This paper proposes an NLP-based optical character recognition-based algorithm, coupled with named entity recognition and hashed database searching to identify and mark pharmaceutical drugs as valid or counterfeit. Optical character recognition using a Tesseract engine would identify the contents of the back-strip of drugs. At the second stage, these contents will pass through a custom named entity tagger to identify the drug name. The third stage will involve the extraction of relevant information from the database to perform a series of validation techniques on the extracted drug data. On processing and confirming these parameters, the drugs will be classified as counterfeit or valid. ...