|
Despite significant strides in diagnosis and treatment, breast cancer remains a formidable health concern, underscoring the ongoing necessity for research. A study employing the polarized Monte Carlo approach has been conducted to investigate and pinpoint breast cancer. Utilizing the Mueller matrix derived from Monte Carlo simulations offers several advantages for diagnostic purposes, including enhanced contrast and the revelation of obscured details. This investigation focuses specifically on scenarios where malignancy is intertwined with normal breast tissue.In addition to its binary classification distinguishing between normal and abnormal conditions, the study presents an additional benefit: pinpointing the center of malignancy in nine specific spatial positions relative to the point of illumination. This offers a means to locate the tumor, even if it is not precisely within the directly illuminated area. The integration of deep learning techniques into a system enables automation and facilitates real-time diagnosis. This research aims to demonstrate the simultaneous detection of both the presence and location of the tumor through Convolutional Neural Network (CNN) implementation on depolarized index images derived from polarized Monte Carlo simulations. The CNN model achieves a classification accuracy of 96%, highlighting its superior performance.
|
