Selected Projects

Here are a selection of projects that I have worked on over the past 2 years.

Contactless Smart Infant Sleep Monitoring System
Sleep Staging

Contactless Smart Infant Sleep Monitoring System

Developed a multispectral physiological imaging system for precise, contactless monitoring of infant vital signs, in collaboration with a leading hospital in Wenzhou. Established an interpretable video-based sleep/wake classification model to enhance monitoring accuracy from consumer-grade to clinical-grade standards. Leveraged ECG and PPG signals to extract and validate infant motion metrics for preliminary sleep analysis. Deployed open-source human pose estimation and optical flow algorithms to compute key-point motion features, confirming feasibility of camera-based polysomnography (PSG) for infant sleep staging. Conducted camera-based PSG monitoring on 100 infants to establish normative sleep-stage benchmarks. Extracted limb movement coordination and intensity features from video data, applying SVM and Random Forest classifiers for binary and multi-class sleep-stage classification. Applied DL algorithms (LSTM, Transformer) to improve the accuracy of infant sleep stage classification.

SSVEP-based BCI Robotic Car Control System with MATLAB
MATLAB-based ECG Signal Acquisition and Processing
Multidimensional Video-based Contactless Infant Seizure Monitoring
Seizure

Multidimensional Video-based Contactless Infant Seizure Monitoring

Developed a real-time monitoring and prediction algorithm for infant seizures in collaboration with a a leading tertiary hospital in Guangzhou, aiming to establish a low-cost, contactless detection system to mitigate resource limitations and inconsistencies in seizure diagnosis quality. Preprocessed raw ECG signals to extract heart rate and calculate heart rate variability (HRV). Utilized remote photoplethysmography (rPPG) to extract heart rate and HRV from video data for non-invasive physiological monitoring. Applied optical flow techniques to analyze global and skin-region motion in vEEG videos. Deployed open-source human pose estimation tools to detect infant keypoints and compute motion intensity. Analyzed limb movement intensity using cross-correlation and Pearson correlation coefficients to integrate motion and physiological signals. Introduced a camera-based solution for NICU settings, enabling contactless monitoring of infant motion and vital signs with improved precision and efficiency.

Research Focus

Overview

Our work bridges computer vision and biomedical engineering to develop non-contact health monitoring technologies, with a primary focus on neonates in the Neonatal Intensive Care Unit (NICU).

Challenge

Traditional monitoring techniques, such as electrocardiography (ECG) and electroencephalography (EEG), pose risks of infection and skin damage, making them unsuitable for long-term use in vulnerable infant populations.

Innovation

To overcome these limitations, we designed a vision-based system employing remote photoplethysmography (rPPG) to estimate vital signs—including heart rate and blood oxygenation—in a completely contact-free manner.

Current Focus

We are expanding this approach to develop an automated sleep staging system for newborns. By integrating multimodal signals such as heart rate, heart rate variability, respiratory rate, and movement patterns captured via camera, the system aims to classify sleep stages autonomously, thereby reducing the operational burden on NICU clinical staff.