Extracting Information from Drug Label Using Image and Text Processing

The purpose of this research aims to develop a method utilizing image and text processing to extract information from drug labels. The input images used in this research are drug labels on the packaging in the form of boxes, blister packs, sachets, or bottles of oral drugs that can be purchased at general drugstores without a prescription, in JPEG, JPG, or PNG format. The extracted information needs to be displayed on the drug label and visible in the image. The process involves image processing, optical character recognition, text processing, and text-to-speech. Experimental results show that the proposed method is effective in extracting information from drug labels with 80% accuracy in drug information searching from drug data file, fair and excellent ranges in information scoring, and 55.6% accuracy of the expiration date retrieval. The best results are obtained from sharp images of drug labels on boxes with standard fonts and clear details. The proposed method can be further developed into an application in the future.