Image processing for correlative light and electron microscopy

People have never stopped exploring the microscopic world. Studying the microstructure of cells helps people better understand the people themselves and has the potential to overcome specific diseases at a fundamental level. Correlative light and electron microscopy (CLEM) can let people intuitively understand the sample information through imaging. In CLEM measurements, samples are measured in both fluorescence microscopy and electron microscopy. Due to technical differences between LM and EM, images obtained from LM and EM contain different information. With the fluorescent labels, one can easily observe the structures of interest. However, due to the diffraction limit, LM image resolution is limited to a few hundred nanometers. EM images can capture the detailed structure of a sample down to the atomic level. However, grayscale images obtained from EM often contain very complex structures. Identifying structures of interest from these complex structures using only EM images is usually a challenge. The CLEM technology provides an opportunity to specify the structures of interest by comparing the CLEM images. However, due to the resolution difference between LM and EM, these structures are usually still not directly distinguishable by simply overlaying the fluorescence microscopy images on high-resolution grayscale electron microscopy images. This thesis aims to investigate a new deconvolution algorithm, EM-guided deconvolution, to automate fusing the LM information on the correlative EM image. We discuss the algorithm with simulated CLEM images and further apply it to experimental data sets. The algorithm can enhance the image resolution to nanometers from correlative wide-field (or confocal) fluorescence microscopy images. The algorithm can effectively recognise, e.g., membrane structures or identify the structures with a suitable point spread function and precise image registration.

Cite

Citation style:
Could not load citation form.

Rights

Use and reproduction: