An Optimal Medicine Identification Application based on Deep Neural Network

Abstract

The paper proposes an advance and enhanced medicine recognition system that assists chronic patients in taking multiple medications based on deep learning techniques, named ST-Med Box. This provides the complete spectrum of functionalities from identifying the right medicine for the patients to giving the timely reminders and maintaining the patient information so that it lessens the chances of wrong intake of drug by the patient that could lead  to harmful interactions of patients with diseases. The proposed system consists of a device that act as an intelligent medicine identification system backed up deep learning module embedded in chipset for training server and its recognition deployment accompanied by an android application that helps to maintain the database management of the patients with its diseases and their medicine in a cloud based management platform and incorporates also with the timely reminders. The experimental result shows that accuracy can be very high and efficient. Therefore, the system designed can help to lessen the risk of incorrect drug interactions and help chronic patients with their treatments in a long-term sustainability as a safe medicational environment reducing the cost expenditure by a chronic patient.