A Novel Mobile Application for Medication Adherence Supervision Based on AR and OpenCV Designed for Elderly Patients

Elderly people or individuals diagnosed in the very early stages of Alzheimer’s Disease might have difficulty in taking medications according to the doctors’ instructions due to the vision problems or memory loss. However, medication adherence can be crucial for good health or cause a major health setback. Traditional apps for medication adherence supervision use time reminders when the doses are supposed to be taken. This paper has developed a novel Android app to track the patient’s action of taking medication for avoiding skipping the doses or taking too much medication. The app includes two modules, i.e. the AR medicine packaging box recognition module and the medication adherence supervision module. The former uses the surface of the medicine packaging box as an AR trigger and displays the virtual words of medical prescription to help the elderly patients identify the correct medications. The second module tracks the date and time of taking medications by recognizing the pills in the photo taken by the user. The pill detection is accomplished by applying the OpenCV image processing library to identify the pill contours based on dimensions and shape. Finally the pill detection tests were conducted to measure the detection accuracy and robustness of the app considering lighting conditions, different locations the pills were placed and the pills overlapped with each other.