The inverse problem resolution of an LED based diffuse reflectance system using artificial neural networks with clinical applications.

This study explores the feasibility of using artificial neural networks (ANN) models trained on reflectance measurements collected from tissue-mimicking liquid phantoms through the DOB Scan Probe to reconstruct topological maps of the reduced scattering coefficient (????????′) and absorption coefficient (????????) that can be used to diagnose cancerous growths and monitor the reduction in tumor sizes for participants undergoing chemotherapy. The first chapter explores breast cancer as a disease and features of the DOB Scan Probe enabling its detection. A review of the various methods available for retrieving ????????′ and ???????? from diffuse reflectance measurements are then provided along with the motivation for their selection. An introduction to ANN models follows, accompanied with a description describing its training. Materials and equipment required to prepare liquid phantoms are listed and the protocol for collecting training data outlined. The thesis ends with a report on the ANN's prediction accuracies for ????????′ and ???????? and a demonstration of its image reconstruction capabilities.