Grants and Contributions:
Grant or Award spanning more than one fiscal year. (2017-2018 to 2022-2023)
Medical diagnosis of malignant melanoma is primarily performed by visual inspection, so that the results can be subjective and variable. In order to standardize the diagnosis practice, my research program focuses on developing inexpensive and easy to operate diagnostic tools for healthcare professionals and the general public to detect melanoma at early stages to improve survival rate. Under my previous NSERC Discovery Grant, I developed a wide range of fully automatic computer programs to detect melanoma-specific pigment disorders from digital skin images, constructed an optic system for measuring skin surface roughness which is another key diagnostic feature used by expert dermatologists to discriminate melanoma from certain benign skin conditions, and implemented a mole tracking computer system which monitors new and disappearing moles over time from photographs of human backs. These research works laid a strong foundation to standardize melanoma detection practice. In this Discovery Grant, I will follow the same research direction, but tackle other technical aspects of computer-aided melanoma diagnosis. In particular, I will investigate following topics:
1) To analyse blood vessel shapes, their arrangements and distributions. These shape features shed light on the diagnosis of melanoma and differentiate it from other common skin conditions.
2) To develop new machine learning techniques for classifying melanoma.
3) To determine appropriate colour parameters for tracing moles over time.
4) To build a new optical device for analysing the polarization property of light. This property could separate melanoma from certain benign skin conditions.
Melanoma is a fatal skin cancer and a growing health problem in Canada and Western Countries. The research results of this proposal will contribute new technologies for building effective diagnostic tools for healthcare professionals and the general public. These tools would help diagnose the disease early and improve the treatment outcomes.