DVTRS- Research University Project

DVTRS which stands for Delineation and Visualization of Tumour & Risk Structures is  one of our major project which has received a substantial funding under the Research University Grant Scheme.

 

Lead Researcher & Project Leader

Prof. Dr. Mandava Rajeswari (Medical Image Analysis)

Co-Researchers

 

  1. Dr. Dhanesh Ramachandram (Medical Image Analysis)
  2. Prof. Dr. Ibrahim Lutfi Shuaib (Consultant Radiologist)
  3. Dr. Mohd Ezane Aziz (Consultant Radiologist)
  4. Assoc.Prof. Dr. Bahari Belaton (Information Visualization)
  5. Dr. Zainul Ahmad Rajion (Consultant Surgeon)
  6. Assoc.Prof. Dr. Sharifah Mastura (Healthcare Information Systems)
  7. Prof. Dr. Rosni Abdullah (Parallel and Distributed Processing)
     

R&D Team

 

  1. Mr. Ong Kok Haur (Msc Candidate + Core Developer)
  2. Mr. Tan Bo (Core Developer)
  3. Mr. Osama Alia (PhD Candidate)
  4. Mr. Mahmoud Jawarneh (PhD Candidate)
  5. Mr. Fariz Ikhwan (Core Developer)
  6. Ms. Anusha Achuthan (PhD Candidate)
  7. Dr.Bong Chin Wei (Post Doctoral Fellow)

 

General Project Objectives

  1. Porting early DVTRS prototype onto an open platform
  2. Enhance quality of 3D visualization
  3. Develop and populate a database of medical cases and images
  4. Research and develop methodologies to reduce manual interaction in delineation in medical images
  5. Research and develop parallelization to speed up processing of medical images
  6. Research on a collaboration framework for medical experts to annotate and share medical image data
  7. Research on several medical image analysis applications
     

 

Overview

 

  1. Research into medical image analysis has been receiving a lot of attention in the past decade. Technology has advanced to the stage where routine medical imaging is performed as the preferred non-invasive approach to diagnosis in many cases. At Universiti Sains Malaysia, research using medical images as the major source of data is being carried out not only at the university’s teaching hospital, but also the newly established Institute for Advanced Medicine and Dentistry at Bertam, as well as at the School of Computer Sciences and to a lesser extent, at the School of Electrical and Electronics Engineering.  Several key areas of expertise that USM excels are drug research, treatment planning, computer vision, medical informatics and information visualization.
  2. The current project is an effort to amalgamate this inter-disciplinary research. As such one of the main deliverables of this research is the software platform for Medical Image Analysis which is being led by the Computer Vision Research Group at the School of Computer Sciences. The software platform, which is currently under active development, would be the basis of several research efforts that involve radiologists, pharmacists, computer scientists and general medical practitioners.

 

 

Project Goals

 

Developing any medical imaging application involve the use of a image laoding, viewing,  preprocessing, segmentation and subsequently analysis. This is achieved through a myriad of proprietary and freely available toolkits. However, integration of these diverse set of development tools is a daunting task.

We hope to provide an extensible and integrated platform for medical image analysis and 3D visualization, incorporating medical image annotation capabilities through an elegant integration of some popula low-level toolkits. The outcome is an medical image viewer and 3D visualizer that is multiplatform, and able to run on typical PC hardware.

Such a tool  could be used by the radiologists to examine medical images and annotate these images, thereby transferring their expertise into a machine readable format, by computer scientists to develop new image analysis algorithms that could be helpful in tissue and tumor delineation as well as assisting in clinical research.

Another key feature that we would like to implement in software development  platform would be the ability to share annotations and expertise between several medical experts. We have already demonstrated this capability through one of our E-Science projects which has been completed albeit on a different software platform.

The software development  platform is being designed to be versatile and extensible using a plug-in architecture which would allow the software to be customized and allows adding or removal of features at any point in time.

While the software platform is largely involves developmental work (programming), the research components will be contributed by several PhD and Msc work which include osteo-sarcoma tumor segmentation, segmentation and quantification of white matter lesions in brain MRI, hippocampus delineation in brain MRI, among others. Another key area of research is the incorporation of knowledge guided segmentation which hopes to take advantage of structured anatomical knowledge (atlas) to produce better segmentation.

Finally, due to the explosive increase in medical image data in hospitals, there is a need to intelligently retrieve relevant medical images which is required by a doctor. This calls for a machine learning approach to aid image and data retrieval.  All these research aspects will be implemented on the  software platform as a form of application plug-ins which could be extended to teh base application we hope to develop all our future research efforts in.

 

Current sub-projects

  1. Extensible Medical Image Analysis and Visualization Platform using Eclipse Plugin Architecture (Endeavor)
  2. Detection and Segmentation White Matter Hyper intensities in MR Images. (MSc)
  3. Assessment of the Influence of Pre-Postoperative Chemotherapy in Osteosarcoma Patients (PhD)
  4. Delineation and Quantification of Hippocampus in Human Brain. (PhD)
  5. Knowledge Guided Medical Image Segmentation(PhD)
  6. Automatic of medical image tagging using machine learning (MSc)