Detecting Breast Cancer from Thermal Infrared Images by Asymmetry Analysis

Abstract

Abstract : This project is a research effort that helps define thermal infrared (IR) imaging as a diagnostic tool in early detection of breast cancer, which can be used as a complementary to traditional mammography. One of the popular methods for breast cancer detection is to make comparisons between contralateral images. In IR imaging, asymmetry analysis normally needs human interference because of the difficulties in automatic segmentation. In order to provide a more objective diagnosis result, we proposed an automatic approach to asymmetry analysis in thermograms. It includes automatic segmentation and pattern classification. The segmentation algorithm uses Hough transform coupled with Canny edge detector to identify four feature curves that define the segments. We propose two pattern classification algorithms, unsupervised clustering and supervised learning based on feature extraction.