Performance evaluation of adaptive meshing algorithms for fluorescence diffuse optical tomography using experimental data

Fluorescence diffuse optical tomography (FDOT) is a computationally demanding imaging problem. The discretizations of FDOT forward and inverse problems pose a trade-off between the accuracy and the computational efficiency of the image reconstruction. To address this trade-off, we analyzed the effect of discretization on the accuracy of FDOT imaging and proposed novel adaptive meshing algorithms for FDOT in a series of studies. In this Letter, we apply these new adaptive meshing algorithms to FDOT imaging using real data from a phantom experiment to demonstrate the practical advantages of our algorithms in FDOT image reconstruction.     

Microfluidic ion-sensing devices

Quantitative determinations of ions in a variety of media have been performed traditionally via one of three approaches: optical instrumental methods (e.g., atomic absorption, and inductively-coupled plasma-optical emission or mass spectrometry), “wet” methods, or ion-selective sensors. Each of the approaches, though, possesses limitations including: power/reagent consumption and lack of portability for instrumental techniques; laborious sample-treatment steps for wet methods; and lack of selectivity and sensitivity with sensors when employed with complex samples. Microfluidic device have emerged as a solution to some of these challenges associated with ion analysis. Such systems can integrate multiple sample handling, calibration, and detection steps (“lab-on-a-chip” concept) into a footprint amenable to portability, while requiring small amounts of sample and power. Furthermore, devices can be constructed for multi-analyte detection, either through multiple parallel fluidic architectures or by using arrays of detection elements. This paper reviews recent progress in the development of total-analysis systems for ionic species. Fabrication techniques and various fluid-handling operations are discussed briefly, followed by a number of more mature strategies for microfluidic ion analysis. A variety of approaches expected to comprise the next generation of devices are also presented.     

Isothermal soot oxidation experiments with intermediate gas change in a Perkin-Elmer TGA6

Two isothermal soot oxidation protocols were tested in a Perkin-Elmer TGA6: (1) the sample was heated under N₂ and then the reaction gas was introduced and (2) the sample was introduced after the empty furnace was heated under the reaction gas. The first protocol is common in soot oxidation studies, it gives a measure of the volatiles and is easier to handle, but as it is shown the determined reaction rate may be falsified by the O₂ concentration. Using gas analysis it was found that ∼2000 s are necessary for the complete gas change in the instrument. An instrument specific correction function involving the O₂ concentration and reaction order n with respect to O₂ was developed which allowed the correlation of the rates measured with both protocols.     

Spectroscopic characterization of micro-phase separated blends of polystyrene/polymethyl methacrylate (PS/PMMA)

An investigation of the miscibility behaviour in polystyrene/ polymethyl methacrylate (PS/PMMA) blends of various compositions under different evaporations protocols using Fourier transform infrared (FT-IR) and Raman Spectroscopic techniques took place in the study. Solvent selection and evaporation rates, coupled with variations in the blend composition resulted in completely different miscibility behaviour for these systems. In particular, it was found that blends with low PMMA content result in systems that exhibit PMMA domains of less than 7 microns on average. Finally, depth profiling studies of the PMMA moiety in the PS matrix show that the distribution of the low content phase is highly affected by the solvent selection as well as the blend composition.     

Coumarin‐containing poly(fluorenediylvinylene)s: Synthesis,photophysics, and electroluminescence

Two new poly(fluorenediylvinylene)s (CV and CF) with coumarin side chains were synthesized via Heck coupling. The coumarin segments were attached to the C‐9 of fluorene through alkyl spacers. The polymers were soluble in common organic solvents such as tetrahydrofuran (THF), chloroform, dichloromethane, and toluene. The photoluminescence (PL), electroluminescence (EL), and electrochemical behavior of these polymers were studied. CV and CF thin films exhibited broad‐band, bluish‐green and orange PL emissions, with maxima at 475 and 585 nm, respectively. These PL maxima were redshifted in comparison with those measured in THF solutions. Aggregate formation played an important role in the solid state. The aggregation was more pronounced in CF thin films than CV thin films. Both polymers oxidized and reduced irreversibly. Light‐emitting devices (LEDs) with indium tin oxide hole‐injecting and aluminum electron‐injecting electrodes were prepared and studied. The LEDs made of CV emitted green light, and the LEDs made of CF exhibited an orange EL emissions. © 2006 Wiley Periodicals, Inc. J Polym Sci Part A: Polym Chem 44: 5750–5762, 2006     

Complete-angle projection diffuse optical tomography by use of early photons

We present the first, to our knowledge, experimental images of complex-shaped phantoms embedded in diffuse media by use of optical tomography. Imaging is based on a complete-angle projection tomographic technique that utilizes transmitted early photons. Results are contrasted with measurements obtained at later gates as well as pseudocontinuous-wave data. The scanning system developed employs noncontact illumination and detection technologies that allow for high spatial sampling of transmitted photons. Combining this system with complete-angle illumination is found to be an important strategy toward improved imaging performance, resulting in a better-posed inversion problem. The appropriateness of reconstruction algorithms similar to those employed in x-ray computed tomography are showcased, and suggestions for model improvements are provided.     

Multispectral optoacoustic tomography by means of normalized spectral ratio

Quantification of biomarkers using multispectral optoacoustic tomography can be challenging due to photon fluence variations with depth and spatially heterogeneous tissue optical properties. Herein we introduce a spectral ratio approach that accounts for photon fluence variations. The performance and imaging improvement achieved with the proposed method is showcased both numerically and experimentally in phantoms and mice.     

Explainable AI: A Review of Machine Learning Interpretability Methods

Recent advances in artificial intelligence (AI) have led to its widespread industrial adoption, with machine learning systems demonstrating superhuman performance in a significant number of tasks. However, this surge in performance, has often been achieved through increased model complexity, turning such systems into “black box” approaches and causing uncertainty regarding the way they operate and, ultimately, the way that they come to decisions. This ambiguity has made it problematic for machine learning systems to be adopted in sensitive yet critical domains, where their value could be immense, such as healthcare. As a result, scientific interest in the field of Explainable Artificial Intelligence (XAI), a field that is concerned with the development of new methods that explain and interpret machine learning models, has been tremendously reignited over recent years. This study focuses on machine learning interpretability methods; more specifically, a literature review and taxonomy of these methods are presented, as well as links to their programming implementations, in the hope that this survey would serve as a reference point for both theorists and practitioners.     

Imager that combines near-infrared diffusive light and ultrasound

We introduce an imaging technique that combines complementary features of ultrasound and near-infrared diffusive light imaging. We achieve the combined technology experimentally by mounting an ultrasound array together with multiple laser source and optical detector fibers upon a hand-held probe. The technique is demonstrated with tissue phantoms wherein both acoustic and optical sensors image the volume underneath the probe. Coregistration of acoustic and optical images is achieved with an accuracy of 0.27+/-0.20cm, approximately half of the image pixel size of our prototype. Accurate determination of target optical absorption is also achieved by use of image segmentation on the ultrasound reconstruction. The combined technique may provide improved breast-cancer detection sensitivity and specificity.