Distinguishing tissue structures via polarization staining images based on different combinations of Mueller matrix polar decomposition parameters

Haoyu Zhai,Yanan Sun,Honghui He,Binguo Chen,Chao He,Yi Wang,Hui Ma
DOI: https://doi.org/10.1016/j.optlaseng.2022.106955
IF: 5.666
2022-05-01
Optics and Lasers in Engineering
Abstract:Mueller matrix polarimetry is gaining wide attention in the field of biophotonics due to its great potential in detecting the microstructures and optical properties of tissue samples label-freely. Individual Mueller matrix derived parameters were often used to characterize certain kind of tissue structure. However, it is found that the individual Mueller matrix parameters only contain partial structural information. Thus, it is difficult to accurately and comprehensively indentify different structures using a single polarization parameter image. Here we introduce an image fusion method based on color spaces to combine different Mueller matrix derived parameters to provide multi-dimensional structural information pixel by pixel in a single polarization staining image. The results of rat back skin tissue specimens indicate that different fibrous structures can be easily distinguished using the polarization staining image. Then, in order to quantitatively analyze the texture characteristics of the polarization staining images for different structures, the Tamura image processing and the gray level co-occurrence matrix (GLCM) methods are adopted to provide various evaluation indices after the images segmentations. The experimental results confirm that the information provided by the polarization staining images based on different Mueller matrix derived parameters can be used for accurate tissue structures discrimination. Combining with machine recognition systems and fast developing artificial intelligence techniques, the strategy proposed in this study can be very helpful for precise abnormal tissues detection and pathological diagnoses.
optics
What problem does this paper attempt to address?