Ganghun Kim

Design and analysis of multi-wavelength diffractive optics.

We present an extension of the direct-binary-search algorithm for designing high-efficiency multi-wavelength diffractive optics that reconstruct in the Fresnel domain. A fast computation method for solving the optimization problem is proposed. Examples of three-wavelength diffractive optics with over 90% diffraction efficiency are presented. These diffractive optical elements reconstruct three distinct image patterns when probed using the design wavelengths. Detailed parametric and sensitivity studies are conducted, which provide insight into the diffractive optics performance when subject to different design conditions as well as common systematic and fabrication errors.

An ultra-small three dimensional computational microscope

We present a computational approach that allows one to create microscopic images in 3D through a miniature, rigid cannula. Light rays propagate from one end of the cannula to the other resulting in a complex intensity distribution. This distribution is unique to the position of the source of the rays. By applying appropriate calibration and utilizing a nonlinear optimization algorithm, we computationally reconstructed images of objects with a minimum feature size of ∼5 μm. Preliminary experiments indicate that the sensitivity of the imaging technique can be as small as 100 nm in the transverse plane and ∼1 μm in the longitudinal direction. Since the cannula is only 14.7 mm long and 200 μm in diameter, this allows for highly miniaturized microscopes that utilize no optics and no scanning. Furthermore, since the images are obtained with just a single frame (no scanning), fast image acquisition is also feasible.

Cannula-based computational fluorescence microscopy

We converted a solid-glass cannula into a high-resolution widefield fluorescence microscope. Calibrating the space-variant point-spread functions of the cannula and applying a nonlinear optimization algorithm to reconstruct object details enable this development. The resolution of our system is ∼1 μm, and fluorophore position is determined to a precision of ∼20 nm. Images of microglia from fixed slices of mouse brains at various post-natal development stages were also obtained.

Numerical analysis of computational-cannula microscopy

Microscopy in hard-to-reach parts of a sample, such as the deep brain, can be enabled by computational-cannula microscopy (CCM), where light is transported from one end to the other end of a solid-glass cannula. Computational methods are applied to unscramble the recorded signal to obtain the object details. Since the cannula itself can be microscopic (∼250  μm in diameter), CCM can enable minimally invasive imaging. Here, we describe a full-scale simulation model for CCM and apply it to not only explore the limits of the technology, but also use it to improve the imaging performance. Specifically, we show that the complexity of the inverse problem to recover CCM images increases with the aspect ratio (length/diameter) of the cannula geometry. We also perform noise tolerance simulations, which indicate that the smaller aspect ratio cannula tolerate noise better than the longer ones. Analysis on noise tolerance using the proposed simulation model showed 2-3× improvement in noise tolerance when the aspect ratio is reduced in half. We can utilize these simulation tools to further improve the performance of CCM and extend the reach of computational microscopy.

Fast imaging in cannula microscope using orthogonal matching pursuit

Fluorescent miscroscopy is a state-of-the-art method for creating high contrast and high resolution images of microscopic structures and has found wide application in microendoscopy (i.e., imaging cellular information from an optical probe within an animal). Cannula based microscopy methods have recently shown great promise for efficient microendoscopy imaging. Yet, performing real-time imaging with cannula methods have yet to be achieved due to the high computational complexity of the algorithms used for image reconstruction. We present an approach based on compressive sensing to improve computational speed and image reconstruction quality. We compare our approach with the state-of-the-art implementation based on direct binary search, a non-linear optimization technique. Results demonstrating up to 70 times improvement in the computation time and visual quality of the image over the direct binary search method are included in the paper.

Deep-brain imaging via epi-fluorescence Computational Cannula Microscopy

Here we demonstrate widefield (field diameter = 200 μm) fluorescence microscopy and video imaging inside the rodent brain at a depth of 2 mm using a simple surgical glass needle (cannula) of diameter 0.22 mm as the primary optical element. The cannula guides excitation light into the brain and the fluorescence signal out of the brain. Concomitant image-processing algorithms are utilized to convert the spatially scrambled images into fluorescent images and video. The small size of the cannula enables minimally invasive imaging, while the long length (>2 mm) allow for deep-brain imaging with no additional complexity in the optical system. Since no scanning is involved, widefield fluorescence video at the native frame rate of the camera can be achieved.

Computational imaging enables a “see-through” lens-less camera

Conventional cameras obscure the scene that is being recorded. Here, we place an image sensor (with no lens) on the edge of a transparent window and form images of the object seen through that window. This is enabled first, by the collection of scattered light by the image sensor, and second, by the solution of an inverse problem that represents the light scattering process. Thereby, we were able to form simple images, and demonstrate a spatial resolution of about 0.1 line-pairs/mm at an object distance of 150mm with depth-of-focus of at least 10mm. We further show imaging of two types of objects: an LED array and a conventional LCD screen. Finally, we also demonstrate color and video imaging.

Computational cannula microscopy: Utilizing a simple glass needle for imaging

Existing brain imaging sensors are either limited in their depth, resolution, and/or inflict trauma. To improve on these limitations, we demonstrate fluorescent microscopy through an optical cannula for deep tissue imaging.

A cannula-based computational fluorescence microscope

We applied nonlinear optimization to convert a rigid cannula into a high-resolution computational-fluorescence microscope. The prototype cannula microscope was used to image fluorescent microspheres and genetically-encoded mouse-brain slices and produced quality images comparable to a conventional widefield microscope.

Fluorescent microscope in a needle

We present a computational microscopy technique to capture fluorescent images through a small glass cannula for minimally invasive in-vivo imaging. Proposed technique produced high fidelity images of microbeads and microglia cells and experimentally achieved resolution up to 1μm.