Michael Chen

EECS Department, University of California, Berkeley

Technical Report No. UCB/EECS-2020-37

May 1, 2020

http://www2.eecs.berkeley.edu/Pubs/TechRpts/2020/EECS-2020-37.pdf

Phase contrast microscopy reveals transparent objects under optical microscopes, and has been widely used for biomedical imaging. Combining specially-designed optics and computational post-processing of the raw acquisitions, quantitative phase imaging (QPI) can convert qualitative phase contrast into physical quantities, such as optical phase delay or refractive index (RI) of the object. A common way to achieve QPI is by creating interference between the scattered light from the object and a reference wave, using coherent illumination. However, the method suffers from low resolution and strong speckle noise, and it is difficult to implement interferometry in off-the-shelf microscopes. In this work, we develop QPI methods for commercial microscopes, with coded illumination. Using a programmable light source (\textit{e.g.} an LED array) that generates partially coherent illumination, many multidimensional QPI techniques can be implemented on a commercially-available microscope without the drawbacks mentioned above. First, differential phase contrast (DPC) microscopy using wavelength-multiplexing is presented, which achieves real-time QPI up to the incoherent resolution limit. Second, quantitative phase and system aberrations are simultaneously encoded in the measurements by alternating between spatially coherent and spatially partial coherent illumination. Absorption and phase of the object, as well as the spatially varying aberrations, can then be resolved computationally. Third, a through-focus DPC microscopy is proposed in order to extend QPI to 3D. The 3D RI of the object is recovered after performing a 3D deconvolution. Finally, most of the existing 3D QPI methods, including 3D DPC, work under the weakly scattering assumption. Hence, they fail to reconstruct accurate RI when multiply scattered light dominates in the captured data. To mitigate artifacts and obtain accurate information of multiple-scattering objects, a new light scattering model is introduced along with an algorithm that solves the non-linear phase retrieval problem. The multiple-scattering model outperforms the traditional weakly scattering models while requiring similar computation cost. As a result, high resolution Giga-voxel 3D phase imaging with multiple-scattering objects is achieved.

Advisors: Laura Waller


BibTeX citation:

@phdthesis{Chen:EECS-2020-37,
    Author= {Chen, Michael},
    Editor= {Waller, Laura},
    Title= {Coded Illumination for Multidimensional Quantitative Phase Imaging},
    School= {EECS Department, University of California, Berkeley},
    Year= {2020},
    Month= {May},
    Url= {http://www2.eecs.berkeley.edu/Pubs/TechRpts/2020/EECS-2020-37.html},
    Number= {UCB/EECS-2020-37},
    Abstract= {Phase contrast microscopy reveals transparent objects under optical microscopes, and has been widely used for biomedical imaging. Combining specially-designed optics and computational post-processing of the raw acquisitions, quantitative phase imaging (QPI) can convert qualitative phase contrast into physical quantities, such as optical phase delay or refractive index (RI) of the object. A common way to achieve QPI is by creating interference between the scattered light from the object and a reference wave, using coherent illumination. However, the method suffers from low resolution and strong speckle noise, and it is difficult to implement interferometry in off-the-shelf microscopes. In this work, we develop QPI methods for commercial microscopes, with coded illumination. Using a programmable light source (\textit{e.g.} an LED array) that generates partially coherent illumination, many multidimensional QPI techniques can be implemented on a commercially-available microscope without the drawbacks mentioned above. First, differential phase contrast (DPC) microscopy using wavelength-multiplexing is presented, which achieves real-time QPI up to the incoherent resolution limit. Second, quantitative phase and system aberrations are simultaneously encoded in the measurements by alternating between spatially coherent and spatially partial coherent illumination. Absorption and phase of the object, as well as the spatially varying aberrations, can then be resolved computationally. Third, a through-focus DPC microscopy is proposed in order to extend QPI to 3D. The 3D RI of the object is recovered after performing a 3D deconvolution. Finally, most of the existing 3D QPI methods, including 3D DPC, work under the weakly scattering assumption. Hence, they fail to reconstruct accurate RI when multiply scattered light dominates in the captured data. To mitigate artifacts and obtain accurate information of multiple-scattering objects, a new light scattering model is introduced along with an algorithm that solves the non-linear phase retrieval problem. The multiple-scattering model outperforms the traditional weakly scattering models while requiring similar computation cost. As a result, high resolution Giga-voxel 3D phase imaging with multiple-scattering objects is achieved.},
}

EndNote citation:

%0 Thesis
%A Chen, Michael 
%E Waller, Laura 
%T Coded Illumination for Multidimensional Quantitative Phase Imaging
%I EECS Department, University of California, Berkeley
%D 2020
%8 May 1
%@ UCB/EECS-2020-37
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/2020/EECS-2020-37.html
%F Chen:EECS-2020-37