NEMA MedBot – AI Posture Analysis System (v2.0) | Axelera AI Challenge 2
Following the recognition received in the first Axelera AI Challenge, we present the evolution of our project: NEMA MedBot, an edge AI system for real-time posture analysis and clinical support.
The system integrates computer vision, natural interaction, and on-device AI inference using Metis 2 and Orange Pi, enabling visual recognition, posture evaluation, and generation of technical reports designed to support specialist assessment. All processing is executed locally, ensuring low latency and data privacy.
The architecture is designed to scale across two complementary solutions: MedBot Desk, a clinical desktop assistant, and Mickey Robot, a mobile cognitive interface for human interaction.
Video demonstration:
https://youtu.be/uzs-fRd8xQ0
GitHub repository:
https://github.com/VITOTRAVE/nema-medbot-edge-demo/tree/main
This work represents an ongoing development toward real-world deployment of AI-based systems supporting healthcare professionals.