Real-Time Medication Identification using Optical Character Recognition

Accurate identification of medication and proper storage are essential to ensure patient safety and preserve drug efficacy. This paper proposes a system titled “Real-Time Medication Optical Character Recognition (OCR) with Storage Requirements”, which combines image-based drug classification with automated retrieval of temperature-specific storage guidelines. A convolutional neural network (CNN) is trained on a public pill image data set to recognize medications based on visual attributes such as color, shape, and imprint. OCR is integrated to extract label text, enhancing the accuracy of the recognition. The identified medication is cross-referenced with a database to display relevant storage instructions in real time. Experimental evaluation using the F1 score, precision recall metrics, and training-validation loss confirms the system’s reliability and effectiveness. The proposed solution offers a practical and scalable approach to improving medication management and storage compliance in clinical and supply chain settings.