Differentiation of mammographic tumor types using the binary joint transform correlator

The performance of the binary joint transform correlator for defining a tumor in a mammographic image as cancerous or benign is investigated and compared to that of the classical correlator. The areas of comparison are the correlation peak and the discrimination sensitivity. The study is a computer simulation of coherent optical processing and includes three separate patient mammograms, two with carcinomas and one with a fibroadenoma. It is shown that the performance of the binary correlator surpassed that of the classical correlator. The binary joint transform correlator is able to differentiate noninvasively and quantitatively between benign and malignant tumors in this experiment