Significant Research Projects (2000-2007)

Visit my research website for a list of my most recent projects.

Adaptive 3D shape reconstruction

A basis function neural network that adaptively determines the control points of a Bézier surface has been developed and successfully applied to the reconstruction of complex bone geometry from serial CT imagery for the manufacture of customized orthopaedic implants. A robust method for registering multiple coordinate data sets for noninvasive measurement of bone kinematics was also investigated. (Collaborator: J.A. Johnson, St. Joseph's Health Centre/UWO)

Data-driven modeling and graphical visualization

Graphical analysis tools provide an environment that can visually enhance patterns embedded in numeric data sets. In this research a spherical self-organizing feature map (SOFM) was used to transform randomly ordered n-dimensional data vectors into content specific colorized shapes that can be manipulated in virtual reality environments. The research was described in a brief article written by a guest editor for "Our Editor Selects: Natural Computing" [Elsevier Mathematics: Volume 2 Issue 4, February 2004]. The basic algorithm of the deformable SOFM has also been applied to scattered data interpolation, free-form surface reconstruction, shape registration, and virtual sculpting. (Collaborators/Virtual Sculpting: R. Canas and D. Johnston, NRC-IMTI)

Shape measurement using an unconstrained range-sensor head

A unique method for measuring complex surface geometry using an unconstrained laser-camera range sensor was developed and patented [US Patent No. 6,542,249]. This initial work has recently lead to a collaborative project with a London-based company to construct a functional hand-held range sensor for developing digital media content. The company creates enterprise class solutions for intelligent media distribution including digital signage and video narrowcasting. A related collaborative project involved the development of a constrained Total-Least-Squares algorithm for computing the transformation parameters of rigid body motion from noisy coordinate data sets. (Collaborator/TLS algorithm: J. Xi, NRC-IMTI/ Ryerson University)

Modeling and analysis of the laser material removal process

Laser micro-machining is a complex nonlinear process with numerous stochastic parameters related to the laser apparatus and the material specimen. This research activity involved the development of a neural network model of the nonlinear laser micro-machining process in an effort to predict the level of pulse energy needed to create a dent with specific depth and diameter. (Collaborators: E.V. Bordatchev and S.K. Nikumb, NRC-IMTI).

Bioelectronic sensors and imaging systems

This multidisciplinary research involves the design and development of high-speed light detectors and imaging arrays that exploit optical and photo-electric properties of thin bacteriorhodopsin (bR) film. Each individual photocell consists of a sandwich-structural device with an ITO (Indium Tin Oxide) electrode/bR film/ITO electrode configuration. A simple fabrication procedure enables a variety sensing array designs to be created on flexible substrates. Preliminary work has resulted in a bioelectronic system for bidirectional motion detection. (Collaborator: A.S. Bassi, Dept. of Chem. & Biochem., UWO).