Postdoctoral Researcher, Department of Biomedical Engineering, McGill University
Affiliated Research Collaborator, Department of Electrical and Computer Systems Engineering, Monash University
Research Field/Interests: Biomedical Signal Processing, Biomedical Health Informatics, Digital Health, Equitable Healthcare, Fetal/Newborn/Child Health
Research Collaborations:
Prof Robert Kearney Lab, McGill University
Perinatal Research Unit, Kaiser Permanente Northern California Division of Research
Prof Yvonne Wu and Dr Marie-Coralie Cornet, University of California, San Francisco
This research focuses on fetal health risk assessment for hypoxic-ischemic encephalopathy (HIE). Past work analysed electronic medical record information of the mother and pregnancy during the antepartum period to identify key risk factors and evaluate HIE risk [paper]. Current work integrates electronic fetal monitoring information (fetal heart rate and uterine contraction) with medical records, maternal vital signs, and laboratory results to generate hourly risk assessment scores. Additionally, ongoing research focuses on developing and validating a labour progress model, which will be integrated into the HIE risk assessment system.
Research Collaborations:
Neo-Beat Lab (Prof Robert Kearney), McGill University
Digital Health Innovation Lab (Prof Guy Dumont) and BC Children's Hospital, University of British Columbia
Biomedical Signal Processing Research Lab (Dr Faezeh Marzbanrad) and Monash Children's Hospital (A/Prof Atul Malhotra), Monash University
This research focuses on automated neonatal health assessment, particularly based on chest sounds [paper1][paper2].
Past research has looked into:
Detection of patent ductus arteriosus [paper]
Cardiac murmur detection and clinical outcome prediction from multi-chest sound recordings from paediatric patients [paper][code]
Respiratory distress prediction at birth from stethoscope recordings of the chest and back [paper]
Explored the effect of surfactant therapy on chest sounds [abstract]
Research Collaborations:
Neo-Beat Lab (Prof Robert Kearney), McGill University
Digital Health Innovation Lab (Prof Guy Dumont) and BC Children's Hospital, University of British Columbia
Biomedical Signal Processing Research Lab (Dr Faezeh Marzbanrad) and Monash Children's Hospital (A/Prof Atul Malhotra), Monash University
This research focuses on using digital stethoscope-acquired chest sounds for health monitoring. This has involved analysing existing digital stethoscopes and the development of newborn-specific digital stethoscopes [abstract].
A crucial step before health assessment is capturing high-quality heart and lung sounds from chest sound recordings. In my past work, I have developed automated signal quality analysis methods for both heart and lung sounds using classical machine learning models, dynamic ensemble classification, and transfer learning [paper1][paper2][article1][article2][code].
Additionally, I have explored denoising and sound separation techniques, particularly leveraging non-negative matrix factorization and non-negative matrix co-factorization to separate chest sound recordings into heart, lung, and noise components [paper1][paper2][paper3][paper4][code].
After ensuring high-quality heart and lung sound acquisition, I have adapted existing methods for neonatal health monitoring, enabling heart segmentation (into S1 and S2 heart sounds), lung segmentation (detecting inspiration and expiration), and estimating heart and breathing rates [paper1][paper2].
More recently, I have begun collaborating on research on measuring newborn milk intake based on acoustic signals [abstract].
Research Collaborations:
Digital Health Innovation Lab (Prof Guy Dumont) and BC Children's Hospital, University of British Columbia
Biomedical Signal Processing Research Lab (Dr Faezeh Marzbanrad) and Monash Children's Hospital (A/Prof Atul Malhotra), Monash University
This research focuses on using videos from RGB cameras for cardiac monitoring with image-based PPG. In my previous work, I implemented neonatal-adapted face and facial landmark detection to identify regions of interest [paper]. I then developed a method for tracking these regions across video frames [paper]. From the tracked regions, I extracted color-based signals to compute PPG and measure heart rate and heart rate variability [thesis].
More recently, I have begun collaborating on research on newborn movement analysis for cerebral palsy assessment.