Ng Kian Wei

I am a full-time data scientist at National University Health System (NUHS) in Singapore. Within NUHS, I manage the technical research and development portfolio for the NUHS Holomedicine Programme, with a focus on mixed reality, applied computer vision and medical image analysis.

Working closely with clinicians and other healthcare professionals, many of the research output have or are undergoing transitions to clinical validation and early stage trials to prepare for wide-scale implementation and usage within hospitals.

On top of my day job as a data scientist, I am also currently a PhD Student at the National University of Singapore, under the supervision of A/Prof Khoo Eng Tat.

About  /  CV  /  Linkedin

Research

We improve on existing mixed-reality ultrasound (MR-US) solutions by developing a portable system that is highly accurate (via stereo triangulation on denser ArUco chessboard keypoints). HoloPOCUS is also the first MR-US work to integrate 3D Ultrasound Reconstruction/Projection using conventional probes.

We utilized depth and infrared sensors on-board HoloLens 2 and developed an all-in-one mixed-reality vein finding and projection system. Spatio-temporal consistency was exploited to provide semi-supervised labels for data-efficient vein segmentation model training.

Teaching

Involved in setting up course materials and provided weekly consultations for a new Data Engineering module catered to engineering undergraduates. Set up and tested data processing cluster on AWS ParallelCluster, integrating Slurm and Bash scripts.

Planned and conducted weekly laboratory sessions, consultations, and grading. Covered introductory programming methodology and basic applications, taught in Python and C. Received above average evaluation and feedback from surveyed students regarding teaching performance.

Media Mentions

"Doctors can soon see scans in real time through an augmented reality (AR) visor that helps them to quickly diagnose injuries or get an update on the status of tumours."

"The veins of some patients can be difficult to find for drawing blood as they may be small or not easily visible under the skin. But artificial intelligence (AI) and augmented reality (AR) may soon address this issue."

Website template courtesy of jonbarron.