Real-time medicine packet recognition system in dispensing medicines for the elderly

Abstract

Currently, Japan is facing an aging society. A larger number of the elderly choose to live in nursing homes, causing the shortage of human resources there. Dispensing medicines is a time and manpower consuming important job in nursing homes. For alleviating the problem, a deep learning-based real-time medicine packet recognition system is designed for assisting dispensing medicines in this paper. The system mainly consists of three stages, including character segmentation by binarization and morphological operations, character recognition by the proposed slight convolutional neural network, and medicine information arrangement such as the patient's name, medicine time. Furthermore, parallel processing is designed for achieving real-time processing. The experimental results show the system realized real-time processing in five fps with 100% accuracy, which demonstrates the effectiveness and practicality of our system.

Keywords

• Medicine packet arrangement
• Dispensing medicines
• Aging society
• Deep learning
• Parallel processing