Research Timeline


CVRG is one of the pioneering research groups that conducted research and development in the field of Computer Vision in Malaysia. Initial research efforts were mainly in general algorithm and technique development. During the boom in semiconductor industry in Penang in the late 80's and through the 90's, considerable research effort in machine vision techniques were put in. Particularly, semiconductor wire bonding inspection, glove-mould inspection and post-mount die attach inspection systems were developed with collaboration from multi-national companies like Intel. Robotics and automation research with an emphasis in intelligent assembly systems was made a focus of research in the late 90's. Robotic vision systems, neural network and force-guided assembly were typical projects at that time. A renewed interest in medical image analysis and image semantics takes us where we are at present.



Current Research

Medical Image Analysis
Medical images have become the de facto modalities used by surgeons and medical practitioners to perform non-invasive examination of the human body. Computed Tomography and MRI are used in conjunction with computer graphics and visualization algorithms to produce an interactive 3D-model of the anatomy of interest. This 3D-model could then be used to facilitate medical experts in diagnosis, surgery and treatment planning. Our focus is on accurate segmentation and quantification of lesions; segmentation and visualization of risk structures including soft and hard tissues, blood vessels etc. We have developed a software tool to address some of the issues and we are interested in extending our research in this area. In addition to building the models we are interested in exploring surgical planning and annotating 3D models with medical ontologies. Some of the immediate projects in this area are:
  • Quantification and visualization of white matter lesions of the brain in a clinical trial of Vitamin E.
  • Assessment of musculoskeletal tumour response to chemotherapy
  • 3D visualization of musculoskeletal tumour and risk structures and surgical planning.
  • Knowledge guided medical image segmentation
  • Segmentation and visualization of blood vessels
Semantic Image Knowledge
The bubble of the information age has burst. The future is in Knowledge. This research is about extracting meaningful knowledge from the sea of information and representing it in a formal way. Image knowledge is knowing or identifying the image contents and their spatial relationships. This understanding represented using a generic terminology that brings same meaning to many, may be considered as knowledge extraction. This generic terminology is referred to as “Ontology”. The extracted knowledge may be represented using semantic networks. Current Image processing techniques are mostly dealing with information at a very low level of the information-knowledge hierarchy whereas human description of images uses concepts that are at the higher level of this heirarchy. The gap between these two representations is known as the “Semantic Gap” problem. Our interest is to address the Semantic Gap with an intention to extract knwoledge about the individual coponents of the image.
  • Image Segmentation
  • Mapping Low-Level Image Features to High-level Semantic Concepts
  • Spatial correlation of concepts
  • Concept disambiguation
  • Multimodal Meaning Normalization Through Ontologies
  • Video annotation and retrieval using semantics


Current Research Projects



  1. 2007-2010      DVTRS - Delineation and 3D Visualization of Tumour and Risk Structures.
  2. 2007-2010      Multimodal Meaning Normalization Through Ontologies.
  3. 2006-2008      Web front-end to the Grid for Resource Sharing, Image analysis and Visualization.
  4. 2007-2010      Narrowing the Semantic Gap : Mapping image features to concepts using ontologies,
  5. 2006-2008      Segmentation and 3D Visualization of Tumor from 2D CT Datasets.
  6. 2005-2007      Hybrid Multi Scale Image Segmentation



Past Research Projects


  1. 2002 -2004     Wavelet-Nurbs Image Segmentation Fundamental Research Grant Scheme.
  2. 2004-2006      Ocular Fundus Image Segmentation for Early Detection of Diabetic Retinopathy.
  3. 2002 -2005     Intelligent Vision Inspection System for Quality Control in Semiconductor Industry
  4. 2000-2002      Shape Similarity Measure Using Polynomial Curve Representation.
  5. 2000-2002      Neural-Fuzzy Force-controlled Robotic Assembly using Virtual prototyping Environment.