Micro-polarimetry for Pre-Clinical Diagnosticsof Pathological Changes in Human Tissues

摘要

The paper presents a practical study of several methods of image analysis applied to polarimetric images of regular andmalignant human tissues. The images of physiological and pathologically changed tissues from body and cervix ofuterus, intestine, kidneys and breast were recorded in transmitted light of different polarization state. The set up of theconventional optical microscope with CCD camera and rotating polarizer's were used for analysis of the polarizationstate of the light transmitted through the tissue slice for each pixel of the camera image. The set of images correspondingto the different coefficients of the Stockes vectors, a 3x3 subset of the Mueller matrix as well as the maps of themagnitude and in-plane direction of the birefringent components in the sample were calculated. Then, the statisticalanalysis and the Fourier transform as well as the autocorrelation methods were used to analyze spatial distribution ofbirefringent elements in the tissue samples. For better recognition of tissue state we proposed a novel method that takesadvantage of multiscale image data decomposition The results were used for selection of the optical characteristics withsignificantly different values for regular and malignant tissues