An Image Processing Based Information Extraction Model for Preventing the Misuse of Drugs

Abstract

Background/Objectives: As the medicine and pharmaceuticals are developed along with the rapid development of the society, human lives are extended to enter the aging society. Accordingly, new drugs are being manufactured and administered, but people in the informational vulnerability class can result in the misuse of drugs due to the absence of information. Therefore, drug information extraction model is required for convenience in verifying the information on drugs.
Methods/Statistical analysis: Non-prescription drugs are being sold worldwide, and the general public is granted with the accessibility and choice on the relevant drugs to increase the awareness on self-medication. People are attempting self-medication to treat their minor symptoms and illnesses, but adverse reactions on drugs from misuse are increasing annually. This is due to the great difficulty in identifying the information on drugs, and therefore in this study, drug images were acquired through sensor elements such as CCD or smartphone camera to identify the information on drugs, and drug shape, character string and color etc. There were analyzed to realize an image processing based drug information model for providing various information on the relevant drugs.
Findings: In this study, an image processing based medical information model was proposed to provide the information on drugs. In the proposed model, images were processed based on ROI setting, image binarization, and histogram equalization technique for quick processing on the inputted images. In addition, a drug information model was realized based on the technologies such as contour algorithm and color identification through RGB decision to extract the accurate shape of the drugs. The proposed model was tested to verify the accurate video image input of the drugs and print of the drug information. It is considered that when the relevant drug is searched through the proposed model in this study, existing inconveniences and difficulties will be reduced, and the existing method of adding various information to search the drug can be deviated to identify the drug information quickly through image acquisition processing.
Improvements/Applications: The proposed model can not only extract drug information, but also be applied to many fields through the recognition of various objects. For the direction of the study in the future, further studies should be continued on the algorithm strong against the changes in the value according to the shade and lighting in the process of acquiring the RGB value on image processing, and on the model on providing the information through the internet website.