We have been developing and applying a new type of polarized light microscope (the P01-Scope) which dramatically enhances the analytic power of the traditional polarizing microscope. The P01-Scope incorporates a universal compensator, made from computer driven liquid crystal devices, to measure the birefringent fine structure at high sensitivity (0.02 nm birefringence retardation), high resolution (0.2 pm), simultaneously over the whole field of view, and for all orientations of the birefringence axis within the focus plane. A 3-dimensional birefringent specimen, such as a living cell, is imaged with the Pol-Scope as a series of optical Sections. Each section contains the birefringence of in-focus structures, such as membranous stacks, aligned filament arrays and individual filament bundles. However, the measurements of in-focus birefringences have spurious contributions from out-of focus cellular components. The structural analysis of the detailed and quantitative information afforded by the P01-Scope is often limited because of these spurious birefringence contributions. Therefore, to take full advantage of the quantitative Pol-Scope images, the out-of-focus information needs to be identified and removed. We are proposing to develop a computational restoration procedure to remove the effect of out-of-focus components on the measurements of in-focus birefringences. The P01-Scope restoration algorithm will be based on procedures established for fluorescence microscopy that we will modify to account for the specific properties of partial coherent imaging with polarized light. The first development goal is to establish a forward procedure that is used to compute image features based on measured point spread functions and known or estimated positions, magnitudes and axis orientations of birefringent objects. The second goal is to find the inverse procedure that optimizes a numerical estimate of object birefringences in volume elements that are contained in P01-Scope optical sections. The inverse procedure iteratively minimizes the difference between experimental and computed images. This process ends in a stable 3-D distribution of positions, magnitudes and axis orientations of specimen birefringences. The image restoration procedures will greatly enhance the utility of the P01-Scope for biological and medical applications to non-invasively analyze the architectural dynamics of living cells.
