Handong Yuan

Monitoring of Blood Pressure Using Non-invasive Sensors

• Developed a non-invasive blood pressure monitoring algorithm using PPG and ECG sensors, enhanced by CNNs and GRUs. Utilised machine learning to improve accuracy and reliability for diverse clinical applications.
• Algorithm achieving Mean Absolute Errors (MAE) of 8.62 mmHg for DBP and 10.69 mmHg for SBP. The majority of the measurement errors were within the 0-5 mmHg range, confirming the model's clinical reliability and precision. Over 94% of the predictions met clinical accuracy standards according to Bland-Altman analysis, underscoring the algorithm's effectiveness in real-world applications.

Smart Mirror Project

• Engineered an interactive Smart Mirror utilizing facial recognition and machine learning to deliver customized affirmations and content, enhancing user engagement in educational settings.
• Programmed and integrated AI functionalities using Python and TensorFlow, focusing on real-time emotional analysis and personalized feedback loops through advanced algorithms.
• Optimized voice interaction features and connectivity to external APIs for dynamic content updates, including weather and calendar events.