Mueller matrix orientation parameters unwrapping for statistically viable distribution

Jiachen Wan,Yue Yao,Yudi Liu,Stephen Arnold,Hui Ma
DOI: https://doi.org/10.1117/12.2577589
2021-01-01
Abstract:Mueller matrix microscopy is a promising non-invasive tool for pathological diagnosis due to its sensitivity to microstructures and its non-reliance on high spatial resolution. Such technique is sensitive to anisotropy, but the majority of such information is deeply hidden within the orientation parameters such as αq, αr, αP and αD. Analysis of them is challenging because orientation parameters varies when the sample’s spatial azimuthal angle changes relative to the imaging system, and the range boundary imposed by the arctan function prevents the parameters from forming a continuous distribution. As the result, the use of orientation parameters is generally avoided during quantitative analysis, despite the rich information they encode. In an effort to resolve these challenges, we propose a novel method for analyzing orientation parameters extracted from Mueller matrix polarimetry. The angular pixel values in the parameter images are unwrapped by assuming continuity, transforming the distorted distribution into one that is statistically viable. The unwrapped orientation parameters are then used for pathological slides analysis. Frequency distribution histograms of the orientation parameters before and after unwrapping are compared, the validity of the proposed method is demonstrated.
What problem does this paper attempt to address?