Search

Browse

• Communities
• Authors
• Year
• Titles
• Subjects

My Space

• Login

Deconvolution microscopy based on modulation transfer function

Boras Charalabos

Full record

---

Summary

Due to the widespread use of the light microscope as a diagnostic tool for many scientific fields like medicine, biology, chemistry as well as for industrial applications, light microscopy has been undergoing vast and continuing innovations both regarding the hardware and software domain. The optical resolution of light microscopes is physically constrained by the phenomenon of diffraction. Out-of-focus light and light originated from adjacent areas of a sample are superposed degrading the quality of the image of the object under study. In this degradation, the Point Spread Function (PSF) of the optical system is the main culprit and it is the one that determines the optical resolution. This degradation effect can be eased by sophisticated and expensive confocal microscopy systems or reversed to some degree by much cheaper widefield deconvolution microscopy methods.Deconvolution processes need a PSF as much as accurate it can be in order to provide satisfactory and realistic results. The description of the PSF can be done either by mathematical models or by experimental measurements. Experiments for this purpose include measurements of fluorescent microbeads as well as estimation of the Modulation Transfer Function (MTF) of the optical systems which finally yields the PSF.However, the conditions of capturing a digital image and the type of noise attributed to the optical systems have led to the creation of many types of deconvolution algorithms that focus and manage these features in a different way. Thus, even a valid PSF of an optical system produces different results when using different algorithm. It is known that the visual perception does not always concur with the quantitative measurements of these processes and ultimately the human eye is making the final decision about the validity of the deconvolution results.This diploma thesis, based on a previous study [28] of the same laboratory and using the experimental methods of extracting PSF and deconvolution, focuses on developing a method that attempts to provide a qualitative comparison of the deconvolution results, which is consistent and verifies visual perception. The method was based on specific variables, such as spatial resolution, noise and image contrast, which were considered to have the greatest contribution to the validity of a result. The final classification list of their algorithms is rankedaccording to their relative successful deconvolution in terms of satisfying these variables.Also, the resulting measurements indicate the behavior of each algorithm in combination with the selected PSF on each sample.This work is integrated into a graphical user interface that provides all the tools for extracting experimental PSFs, implementing a range of iterative and non iterative, deconvolutionalgorithms, qualitative analysis of deconvolution results to create ranking list and export statistics for creating completed reports in each experiment.

Available Files

Services