Research and Publications 

Medical imaging has produced a large number of volumetric images capturing biological structures in 3D. Computer-based understanding of these structures can often benefit from the knowledge of shape components, particularly rod-like and plate-like parts, in such volumes. Previously, skeletons have been a common tool for identifying these shape components in a solid object. However, obtaining skeletons of a grayscale volume poses new challenges due to the lack of a clear boundary between object and background. In this paper, we present a new skeletonization algorithm on grayscale volumes typical to medical imaging (e.g., MRI, CT and EM scans), for the purpose of identifying shape components. Our algorithm does not require an explicit segmentation of the volume into object and background, and is capable of producing skeletal curves and surfaces that lie centered at rod-shaped and plate-shaped parts in the grayscale volume. Our method is demonstrated on both synthetic and medical data.

In this paper, we describe a novel geometric approach in the process of recovering 3D protein structures from scalar volumes. The input to our method is a sequence of alpha-helices that make up a protein, and a low-resolution protein density volume where possible locations of alpha-helices have been detected. Our task is to identify the correspondence between the two sets of helices, which will shed light on how the protein folds in space. The central theme of our approach is to cast the correspondence problem as that of shape matching between the 3D volume and the 1D sequence. We model both shapes as attributed relational graphs, and formulate a constrained inexact graph matching problem. To compute the matching, we developed an optimal algorithm based on the A*-search with several choices of heuristic functions. As demonstrated in a suite of synthetic and authentic inputs, the shape-modeling approach is capable of identifying helix correspondences in noise-abundant volumes at high accuracy with minimal or no user intervention.

The rapid growth in global mobile phone usage has created a great demand for feature rich mobile applications. However, the resource limitations in mobile devices and the slow, unreliable nature of mobile networks are two major hindrances to the provision of such services. This paper presents a generic framework for mobile application development, which provides guidelines to develop an application as a collection of independently executable segments. Further, an execution mechanism is proposed which enables these code segments to be downloaded “on demand”. A network infrastructure using 3G mobile technologies is used to mitigate the problem of low bandwidth that may hinder the timely transfer of segments. The authors conclude that the concept of “segmented application development” and “code-on-demand” discussed in this paper could change the nature and quality of applications available for resource constrained mobile devices.

Today, three-dimensional motion tracking is implemented using magnetic, fibre optic and mechanical techniques that share a common setback: the need for physical contact with the target. Although vision-based techniques provide a contact-free motion tracking solution, they have not been commercially used due to their high resource requirements and code complexity. This paper describes a generic platform that would hide the complexity of vision-based techniques and provide location information via a simple and open protocol. The first step of the solution involves capturing information using multiple image sources, which can be low-cost web-cams or even specialised wide-angle cameras. These image streams are thereafter sent to the Server component that separates the targets from the background using image differencing and a threshold function. Thereafter, a noise reduction algorithm is used to eliminate salt and pepper noise. The flood-fill algorithm is used on the result to identify the borders of each target within each image stream. Finally, the three-dimensional locations of the targets are calculated within the server component using Epipolar geometry. The location information is thereafter sent to software and hardware clients using an open protocol based on the Extensible Markup Language. The RSA encryption algorithm is used in this protocol to ensure the confidentiality of the information being transmitted. Analysis of the developed prototype has demonstrated its practical applicability thereby making vision-based three-dimensional motion tracking more accessible to the commercial and academic worlds.

• Gold Medal, National Best Quality Software Awards, 2004
• Sri Lankan Nomination for the Asia Pacific ICT Awards, 2004