Deep Compressive Microendoscopy Technology

Date

2024-8-20

Author

Çılgın, Zafer

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

166
views

0
downloads

Cite This

Staining sparsely populated cells and examining them under light microscopy has dramatically transformed our understanding of the nervous system. Today, far-field optical imaging microscopy is an indispensable tool for neuroscientists seeking to uncover the mechanisms underlying nervous system dynamics. In conventional far-field optical imaging, light diffraction limits spatial resolution, while the Nyquist–Shannon sampling theorem constrains temporal resolution. Using optical, computational, and statistical techniques, super-resolution methods applied to living cells have surpassed the diffraction limit and enhanced spatial resolution. However, these techniques often require high-power illumination, which can damage cells, causing photo-toxicity and photo-bleaching. Also, super-resolution methods are time-consuming and struggle to balance spatial and temporal resolution effectively. Improving temporal resolution remains a priority for researchers in this field. Compressed Sensing (CS) is a novel theory that reconstructs signals using less data than traditional methods. While CS reduces data measurements in microendoscopy, its performance declines due to the sparsity assumption in data-intensive applications. Advances in computer science, particularly Deep Learning, have facilitated the automation of optical image reconstruction, eliminating time-consuming, harmful, and costly processes. Given these advancements, research that enables high temporal resolution optical image reconstruction while avoiding destructive and costly methods will significantly contribute to the field. Deep Compressive Microendoscopy Technology (DCMT) aims to achieve this by leveraging extensive raw datasets to understand the physics of imaging systems and performing compressed sensing optical imaging. This approach promises a healthier and more efficient way of transmitting information, minimizing the impact of optical elements.

Subject Keywords

Compressed Sensing, Deep Learning, Lensless Imaging, Microendoscopy, Speckle-Based Imaging

URI

https://hdl.handle.net/11511/111036

Collections

Graduate School of Natural and Applied Sciences, Thesis