MediAI: Intelligent AI System for Medicine Identification, Symptom Analysis Chatbot, and QR Code Tracking for Central Province, Sri Lanka

Abstract

Medication errors and poor compliance remain major global healthcare challenges, particularly in Sri Lanka, where nearly 40% of patients struggle to identify their medications. These issues contribute to hospital readmissions, financial burdens, and preventable complications. To address this, we present MediAI, an AI-powered medication management system designed to improve adherence and reduce errors. MediAI combines three key components: a deep learning-based medicine identification model using EfficientNet, a generative AI chatbot for symptom analysis, and a QR code-based prescription tracking system. The model was trained on a custom dataset of over 2,700 high-resolution drug images and achieved a classification accuracy of 96.87%. A mixed-methods evaluation—including surveys, interviews, and pilot testing in hospitals and pharmacies—revealed significant improvements in prescription verification, patient compliance, and medication tracking. While MediAI has demonstrated strong potential, current limitations include dependency on internet access and limited handling of complex medical queries. Future development will focus on expanding the image dataset, introducing multi-language support, and enabling offline access to better serve rural and underserved communities. MediAI offers a scalable, impactful solution that bridges traditional and intelligent medication management, aiming to enhance patient safety and healthcare efficiency in low-resource settings.

Ähnliche Arbeiten