Masterprojekt

MediScan-App

 

The MediScan app for medication management provides essential information from a medication leaflet in a simple and structured format. Through digital processing, the information becomes quickly and clearly accessible to users. Additional features in the mobile application enable users to create clear medication plans.

Objectives

  • Development of a medication scanning app
  • Medication identification by approval number with a smartphone camera
  • Extraction and clear presentation of leaflet information
  • Personalization through the creation of a user profile
  • Digital medication cabinet with a dosing schedule

Technical Background

To develop the app, it was divided into two working areas (frontend and backend). In the frontend, which creates the user interface, Flutter was used for programming, while in the backend, Natural Language Processing (NLP) was developed using Python. The NLP is based on OpenAI's GPT 3.5 and is responsible for extracting specific information from the package inserts.

Dataflow of the MediScan-App. © FH JOANNEUM / eHealth

Results

The outcomes of the project include a functional mobile application that facilitates the handling of medication information for users, as well as an AI-driven NLP logic that extracts relevant information from the medication package inserts and presents it in a user-friendly manner.

MediScan-App 3

Screenshots of the MediScan-App: drug scanner, medical cabinet and drug information. © FH JOANNEUM / eHealth

MediScan-App 2

Project team (from left to right: Oliver Perkics, Lisa Heiler, Katharina Kirchsteiger, Simon Hartl) during the poster walk. © FH JOANNEUM / eHealth