Publication     Conference     Thesis
Resources      Demo

Research Highlight

Skull Reconstruction

Human skulls are the target objects in craniofacial surgery planning, forensic investigation and physical anthropological study. In cramofacial surgery planning, a patient's skull can be defective due to impact injury, congenital deformity or tumor. To repair the patient's skull accurately, a surgeon requires a normal reference model to guide his operations. In forensic investigation, a victim's skull can be defective due to impact injury caused by violence. To identify the victim, an investigator needs to reconstruct the victim's skull followed by facial reconstruction. In physical anthropological study, a human or non-human hominid skull can be defective in many ways including natural causes. To fully characterize the skulls, an anthropologist needs to reconstruc the subject's skull from the defectivfe ones.

Our research goal is to develop software tools and algorithms for reconstruction of normal skulls from defective skulls that are applicable to craniofacial surgery, forensic investigation and physical anthropology.

This project is collaborated with National University Hospital.


Medical Image Analysis
 
  Skull Reconstruction

This project aims at developing software tools and algorithms for reconstruction of normal skulls from defective skulls that are applicable to craniofacial surgery, forensic investigation and physical anthropology.

This project is collaborated with National University Hospital.
     
  Stroke Analysis

This project aims at developing software tools for the quantitatiive analysis of stroke patient's CTA images to assist in diagnosis.

This project is collaborated with National University Hospital.
     
  Simulation and Planning of Cardiovascular Surgery

This project aims at developing a system for the simulation and planning of cardiovascular surgery that involves major blood vessels such as the aorta, pulmonary artery, vena cava, and coronary arteries.

This project is collaborated with National Taiwan University Hospital.
     
  Analysis of Facial Images for Acne Classification and Quantification

The objective of this project is to develop a system for classifying the types of acnes on a patient's face and to quantifying them (count number, compute area, etc.).

This project is collaborated with National Skin Centre.
     
  MMedWeb – Multimedia Medical Conceptual Web for Intelligent Information Access

This project aims at developing a framework for the semantic structuring and organization of multimedia medical information, including medical images and medical reports.

This is a collaborative project with I2R, CNRS, France; Singapore General Hospital, and National University Hospital.
     
  Model-Based Segmentation of Medical Images

This project's objective is to develop model-based methods for segmenting anatomical structures in complex medical images.

This is a collaborative project with Faculty of Medicine, NUS, and National University Hospital, and Singapore General Hospital.
     
  X-ray Fracture Detection

The objective of this project is to develop a system for screening and detection of fractures in x-ray images of femur, wrist, etc.

This project is collaborated with Singapore General Hospital. We are now developing a prototype system for field test in Singapore General Hospital.
     
Computer Vision
 
  Robust Principal Component Analysis

Our research goal is to advance the research of low-rank matrix recovery for visual computing by overcoming the above obstacles so that a unified approach can be applied to solve a wide range of problems in various application domains.
     
  3D Motion Capture for Digital Cultural Heritage

The objective is to apply digital media technologies to bring cultral heritage to live!
     
  3D Motion Analysis for Sports Coaching

This project's objective is to develop a system for sports coaching by matching 2D motion of a novice with 3D motion of an expert. Target sports include short burst sucha as golf swing and long, complex motion such as Taichi.

> Demo:
      input-1, output-1
      input-2, output-2
      input-3, output-3
     
  City-Wide Surveillance with Multiple Cameras

Our research goal is to develop a framework for city-wide surveillance, which includes tracking moving objects of interests over multiple camera views, analyzing the motion paths, and recognizing abnormal motion and behavior.
     
  Rain Removal

The objective of this project is to develop an algorithm that detects rain drops in a video clip and remove them. The algorithm has been applied to videos taken with both stationary and moving cameras.

> demo
     
  Perceptual Texture Space

This project constructs a perceptual texture space in which texture similarity is measured in a perceptually consistent manner. The perceptual texture space is invariant to intensity, contrast, scale, orientation.
     
Other Research
 
  Protein Docking

This project's objective is to develop an algorithm that predicts possible binding of two proteins based on their 3D structures and electrostatic and chemical properties.
     
  Robot Motion Planning and Control
Multi-Robot Task Allocation


This project studies motion planning of multiple robots and the dynamic allocation of tasks to the robots in response to chaninging environmental inputs.

8 Nov 2019