All dielectric metasurface of low loss TiO<sub>2</sub> in visible wavelengths was devised forming subwavelength-scale nanostructures. DC-magnetron sputtering of oxygen-reduced TiOx (x<;2) target with reactive oxygen gas made dense amorphous TiO<sub>2</sub> layers of 1.35nm Ra. A 380nm-thick TiO<sub>2</sub> has low extinction coefficient (k) under 1x10<sup>-5</sup>, transparency of 98.33% and high refractive index (n) of 2.55 at 485nm wavelength. Highly precise TiO<sub>2</sub> meta-atoms were successfully defined with 100nm-300nm feature size. Phase shift properties of TiO<sub>2</sub> metasurface were measured. Finally, we constructed dielectric metaphotonic platform for various optical devices such as band pass filters, flat lens and beam deflectors in visible ranges.
Multiple light scattering in tissue limits the penetration of optical coherence tomography (OCT) imaging. Here, we present in vivo OCT imaging of a live mouse using wavefront shaping (WS) to enhance the penetration depth. A digital micromirror device was used in a spectral-domain OCT system for complex WS of an incident beam which resulted in the optimal delivery of light energy into deep tissue. Ex vivo imaging of chicken breasts and mouse ear tissues showed enhancements in the strength of the image signals and the penetration depth, and in vivo imaging of the tail of a live mouse provided a multilayered structure inside the tissue.
This paper describes a novel quantitative phase microscopy based on a simple self-referencing scheme using
Michelson interferometry. In order to achieve the homogeneous reference field for accurate phase measurement,
the imaging field-of-view (FOV) is split onto the sample and homogenous background areas. The reference field
can be generated by rotating the relative position of the sample and homogenous background in the object arm.
Furthermore, our system is realized using an extended depth-of-field (eDOF) optics, which allows quantitative
phase measurement for an increase of the depth-of-field without moving objective lens or specimen. The proposed
method is confirmed by experimental results using various samples such as polystyrene beads and red blood cells
Near-infrared spectroscopy (NIRS) can be employed to investigate brain activities associated with regional changes of the oxy- and deoxyhemoglobin concentration by measuring the absorption of near-infrared light through the intact skull. NIRS is regarded as a promising neuroimaging modality thanks to its excellent temporal resolution and flexibility for routine monitoring. Recently, the general linear model (GLM), which is a standard method for functional MRI (fMRI) analysis, has been employed for quantitative analysis of NIRS data. However, the GLM often fails in NIRS when there exists an unknown global trend due to breathing, cardiac, vasomotion, or other experimental errors. We propose a wavelet minimum description length (Wavelet-MDL) detrending algorithm to overcome this problem. Specifically, the wavelet transform is applied to decompose NIRS measurements into global trends, hemodynamic signals, and uncorrelated noise components at distinct scales. The minimum description length (MDL) principle plays an important role in preventing over- or underfitting and facilitates optimal model order selection for the global trend estimate. Experimental results demonstrate that the new detrending algorithm outperforms the conventional approaches.
Proc. SPIE. 6850, Multimodal Biomedical Imaging III
KEYWORDS: Data modeling, Magnetic resonance imaging, Wavelets, Linear filtering, Near infrared spectroscopy, Brain activation, Neuroimaging, Positron emission tomography, Functional magnetic resonance imaging, Brain
Near infrared spectroscopy (NIRS) is a relatively new non-invasive brain imaging method to measure brain
activities associated with regional changes of the oxy- and deoxy- hemoglobin concentration. Typically, functional
MRI or PET data are analyzed using the general linear model (GLM), in which measurements are modeled as a
linear combination of explanatory variables plus an error term. However, the GLM often fails in NIRS if there
exists an unknown global trend due to breathing, cardiac, vaso- motion and other experimental errors. In order
to overcome these problems, we propose a wavelet-MDL based detrending algorithm. Specifically, the wavelet
transform is applied to NIRS measurements to decompose them into global trends, signals and uncorrelated
noise components in distinct scales. In order to prevent the over-fitting the minimum length description (MDL)
principle is applied. Experimental results demonstrate that the new detrending algorithm outperforms the
Proc. SPIE. 6850, Multimodal Biomedical Imaging III
KEYWORDS: Statistical analysis, Data modeling, Magnetic resonance imaging, 3D modeling, Smoothing, Near infrared spectroscopy, Scanning probe microscopy, Brain mapping, Functional magnetic resonance imaging, Brain
Even though there exists a powerful statistical parametric mapping (SPM) tool for fMRI, similar public domain
tools are not available for near infrared spectroscopy (NIRS). In this paper, we describe a new public domain
statistical toolbox called NIRS-SPM for quantitative analysis of NIRS signals. Specifically, NIRS-SPM statistically
analyzes the NIRS data using GLM and makes inference as the excursion probability which comes from
the random field that are interpolated from the sparse measurement. In order to obtain correct inference, NIRS-SPM
offers the pre-coloring and pre-whitening method for temporal correlation estimation. For simultaneous
recording NIRS signal with fMRI, the spatial mapping between fMRI image and real coordinate in 3-D digitizer
is estimated using Horn's algorithm. These powerful tools allows us the super-resolution localization of the brain
activation which is not possible using the conventional NIRS analysis tools.
This paper describes a novel reconstruction algorithm for microscopy axial tomography, which reconstructs a
3-D volume using multiple tilted views through an off-centered aperture and numerical processing. The main
contribution of this paper is a derivation of novel optimization criterion and algorithm for a cost function
with <i>L</i><sub>1</sub> fidelity term and sparsity constraint. A parallel coordinate descent (PCD) algorithm has been derived
as an efficient optimization methods, which corresponds to iterative application of projection and <i>nonlinear</i>
back-projection using median. Numerical simulation results using synthetic and real microscopy data show
that accurate reconstruction can be obtained rapidly, and interference artifacts from high contrast objects in a
volume can be removed efficiently. Our algorithm is quite general, and can be used for many other tomosynthesis
applications with limited number of views.