Characterization and three-dimensional localization of cancerous prostate tissue using backscattering scanning polarization imaging and independent component analysis
Image Interpretation, Computer-Assisted
Pattern Recognition, Automated
Characterization and three-dimensional (3-D) localization of human cancerous prostate tissue embedded in normal prostate tissue were demonstrated using backscattering scanning polarization imaging and an inverse imaging reconstruction algorithm, optical tomography using independent component analysis (OPTICA). Two-dimensional (2-D) backscattering images of a prostate tissue sample illuminated with a scanning laser beam were measured with a CCD camera to obtain multiple angular views of the target embedded inside the tissue. The recorded sets of 2-D images were used to determine the existence and 3-D location of the cancerous prostate tissue using the algorithm. The difficulty arises in the backscattering geometry because the profile of the incident beam and the surface property of the tissue sample appreciably affect the spatial distribution of the backscattered light. This challenge was addressed by: (1) synthesizing a "clean" background image of the host medium; and (2) numerically marching the propagation of the scattered light from the hidden target to the surface of the tissue sample until matching the retrieved independent component. The OPTICA algorithm was improved specifically for the backscattering model, and used to obtain 3-D locations of the cancerous tissue embedded in normal host tissue. The retrieved results were found in good agreement with the known 3-D positions of the cancerous tissue.