Image correction algorithm for functional three-dimensional diffuse optical tomography brain imaging

We outline a computationally efficient image correction algorithm, which we have applied to diffuse optical tomography (DOT) image time series derived from a magnetic resonance imaging (MRI)-based brain model. Results show that the algorithm increases spatial resolution, decreases spatial bias, and only modestly reduces temporal accuracy for noise levels typically seen in experiment, and produces results comparable to image reconstructions that incorporate information from MRI priors. We demonstrate that this algorithm has robust performance in the presence of noise, background heterogeneity, irregular external and internal boundaries, and error in the initial guess. However, the algorithm introduces artifacts when the absorption and scattering coefficients of the reference medium are overestimated--a situation that is easily avoided in practice. The considered algorithm offers a practical approach to improving the quality of images from time-series DOT, even without the use of MRI priors.     

Relating aviation service difficulty reports to accident data for safety trend prediction

Commercial aviation fatalities are predicted (Fulwood et al., ‘Relating aviation service difficulty reports to accident data for safety trend prediction’, BNL Tech. Rpt 63018, March 1996) using linear superposition of the time-dependent spectra of key aircraft systems difficulties reporting in FAA's Service Difficulty Reports (SDR (DOT, FAA, ‘Flight standards service difficulty program’, Order 8010.2, Feb. 22, 1978, reissued 4:5/14/81)) data base. The fitting coefficients are found by a linear regression model (referred to as ‘the model’) to FAA's Accident Incident Data System (AIDS (DOT/FAA, ‘Aviation standards accident/incident data system—AIDS’, Users Guide VS ASAS-D-335, July 1982)) covering 5.5 years beginning January 1990. The model was tested by dividing the data approximately in half, using the first half to calibrate the model for prediction of the second half. A second test did the opposite. A third test used the first 60 months of data to predict the following 6 months. These tests (Fig. 5) showed good agreement between the model and AIDS data.The deficiency frequency of ATA (Aircraft Transportation Association) systems is reported (Table 2). Third-order fitting of the AIDS data was also used for prediction. All methods are compared in Fig. 4. The model was deemed superior because it reflects inspections and may be updated with SDR data from the Internet. The magnitude of the model's fitting coefficients indicate systems importances to the results.     

Feed‐forward neural network optimized by hybridization of PSO and ABC for abnormal brain detection

Automated and accurate classification of MR brain images is of crucially importance for medical analysis and interpretation. We proposed a novel automatic classification system based on particle swarm optimization (PSO) and artificial bee colony (ABC), with the aim of distinguishing abnormal brains from normal brains in MRI scanning. The proposed method used stationary wavelet transform (SWT) to extract features from MR brain images. SWT is translation‐invariant and performed well even the image suffered from slight translation. Next, principal component analysis (PCA) was harnessed to reduce the SWT coefficients. Based on three different hybridization methods of PSO and ABC, we proposed three new variants of feed‐forward neural network (FNN), consisting of IABAP‐FNN, ABC‐SPSO‐FNN, and HPA‐FNN. The 10 runs of K‐fold cross validation result showed the proposed HPA‐FNN was superior to not only other two proposed classifiers but also existing state‐of‐the‐art methods in terms of classification accuracy. In addition, the method achieved perfect classification on Dataset‐66 and Dataset‐160. For Dataset‐255, the 10 repetition achieved average sensitivity of 99.37%, average specificity of 100.00%, average precision of 100.00%, and average accuracy of 99.45%. The offline learning cost 219.077 s for Dataset‐255, and merely 0.016 s for online prediction. Thus, the proposed SWT + PCA + HPA‐FNN method excelled existing methods. It can be applied to practical use.     

Pediatric brain extraction from T2-weighted MR images using 3D dual frame U-net and human connectome database

Accurate extraction of brain tissues from magnetic resonance (MR) images is important in neuroradiology. However, brain extraction is more difficult for pediatric brains than for adult brains due to several factors including smaller brain sizes and lower tissue contrasts. In this work, we propose a brain extraction technique that utilizes dual frame (DF) 3D U-net deep learning architecture and the human connectome project (HCP) database for multislice 2D pediatric T2-weighted MR images with diseases. To improve segmentation accuracy in small pediatric brains with detailed boundary regions, DF 3D U-net architecture was used. We pretrained networks with the HCP database to compensate for the limited amount of MR images and manual segmentation masks of pediatric patients. For quantitative analysis, we compared the brain extraction results of brain extraction tool, DF, and conventional 3D U-net using the dice similarity coefficient (DSC), intersection of union (IoU), and boundary F1 (BF) scores; each deep learning architecture was evaluated with and without pretraining using the HCP. This study included 10 patients with diseases and all images were acquired using a PROPELLER MR sequence. Pretraining using the HCP database enhanced segmentation performance of the network, and the skip connections in the DF 3D U-net could enhance the contour similarity of segmentation results. Experimental results showed that the proposed method increased the DSC, IoU, and BF scores by 0.8%, 1.6%, and 1.5%, respectively, compared with those of the conventional 3D U-net without pretraining.     

Measurements of the foreign-broadened continuum of water vapor in the 6.3 microm band at -30 degrees C

The foreign-broadened continuum of water vapor in the nu2 band (5-7.7 microm, 1300-2000 cm(-1)) is important for satellite-based retrievals of water vapor in the upper troposphere, where temperatures are below -25 degrees C. Continuum coefficients have previously been measured mostly at or above +23 degrees C. We present continuum coefficients in the nu(2) band retrieved from measurements made in Antarctica at temperatures near -30 degrees C: atmospheric transmission at South Pole Station and atmospheric emission at Dome C. The continuum coefficients derived from these measurements are generally in agreement with the widely used Mlawer, Tobin-Clough, Kneizys, Davies continuum. Differences are at most 30%, corresponding to a 6% relative error in retrieved upper-tropospheric humidity.     

Dependence of GMR on NiFe layer thickness in high sensitive simple spin valve

The dependence of the giant magnetoresistance on Ni/sub 81/Fe/sub 19/ soft magnetic layer thickness is investigated experimentally for a simple spin valve with a top-pinned structure of Ta (6 nm)/Ni/sub 81/Fe/sub 19//Co/sub 90/Fe/sub 10/ (1 nm)/Cu (1.8 nm)/Co/sub 90/Fe/sub 10/ (3.5 nm)/Ir/sub 20/Mn/sub 80/ (8 nm)/Ta (6 nm). With Ni/sub 81/Fe/sub 19/ thickness increased from 6 nm to 7 nm, the magnetoresistance (MR) ratio decreases sharply from 8.34% to 3.34%, whereas it changes only slightly within the thickness ranges from 2-6 nm and from 7-12 nm, and larger MR ratios are obtained in the range from 2-6 nm. For a spin valve with an optimized thickness of Ir/sub 20/Mn/sub 80/ (11 nm) and top Ta (3 nm), the MR dependence is in accordance with the former structure when Ni/sub 81/Fe/sub 19/ thickness changes from 3.5 to 5.5 nm, and an optimized spin valve with 4.5-nm-thick Ni/sub 81/Fe/sub 19/ is obtained. This spin valve has a large MR ratio (9.15%), low coercive force (0.85 Oe), and high sensitivity, which makes it promising for applications.     

Improved accuracy of reconstructed diffuse optical tomographic images by means of spatial deconvolution: two-dimensional quantitative characterization

Systematic characterization studies are presented, relating to a previously reported spatial deconvolution operation that seeks to compensate for the information-blurring property of first-order perturbation algorithms for diffuse optical tomography (DOT) image reconstruction. In simulation results that are presented, this deconvolution operation has been applied to two-dimensional DOT images reconstructed by solving a first-order perturbation equation. Under study was the effect on algorithm performance of control parameters in the measurement (number and spatial distribution of sources and detectors, presence of noise, and presence of systematic error), target (medium shape; and number, location, size, and contrast of inclusions), and computational (number of finite-element-method mesh nodes, length of filter-generating linear system, among others) parameter spaces associated with computation and the use of the deconvolution operators. Substantial improvements in reconstructed image quality, in terms of recovered inclusion location, size, and contrast, are found in all cases. A finding of practical importance is that the method is robust to appreciable differences between the optical coefficients of the media used for filter generation and those of the target media to which the filters are subsequently applied.     

Dynamic estimation of the modeling error statistics in Diffuse Optical Tomography

Diffuse Optical Tomography (DOT) is a non-invasive imaging technique that suffers from a typical large-scale and ill-posed inverse problem with low spatial resolution. In DOT, the inverse problem is computationally intensive and decreasing the computation complexity and making it well-posed is the one of the most challenging research areas. More precisely, one of the well-known complexity reduction techniques is defined as applying modelling error originated from discretization of forward problem. Applying the discretization error in Bayesian inference has already been discussed; the method in which the likelihood is modified by an off-line prior density estimation. This paper implements a new method to enhance the modelling error approach using an iterative scheme to update statistical parameters of modelling discrepancy in DOT. The algorithm is very similar to Ensemble Kalman Filter. Moreover, the reconstruction process in the applied method is conducted by a small sample size rather than off-line method. Hence, the computation complexity is decreased and the algorithm converges in few iterations. The efficiency of the proposed method is illustrated by simulations.     

Hierarchical fuzzy segmentation of brain MR images

In brain magnetic resonance (MR) images, image segmentation and 3D visualization are very useful tools for the diagnosis of abnormalities. Segmentation of white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF) is the basic process for 3D visualization of brain MR images. Of the many algorithms, the fuzzy c‐means (FCM) technique has been widely used for segmentation of brain MR images. However, the FCM technique does not yield sufficient results under radio frequency (RF) nonuniformity. We propose a hierarchical FCM (HFCM), which provides good segmentation results under RF nonuniformity and does not require any parameter setting. We also generate Talairach templates of the brain that are deformed to 3D brain MR images. Using the deformed templates, only the cerebrum region is extracted from the 3D brain MR images. Then, the proposed HFCM partitions the cerebrum region into WM, GM, and CSF. © 2003 Wiley Periodicals, Inc. Int J Imaging Syst Technol 13, 115–125, 2003; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ima.10035     

Evaluation of test procedures for determining servo compatibilityof heads and media in magnetic disk drives

We evaluate a testing specification proposed by the National Storage Industry Consortium's (NSIC) Extremely High Density Recording (EHDR) group for evaluating head and media compatibility for servo performance in magnetic disk drives. These tests use average amplitude and average noise profile measurements across isolated tracks to predict the shape, linearity, noise, and long-term stability of position error signal (PES) patterns. We compare the predictions from these tests to measurements from null and amplitude PES patterns written on a spin-stand. Results show average PES-profile prediction errors of 1%-2% track width and noise level prediction within a factor of 2. We present data from tests for long-term stability of the magnetoresistive (MR) read element after repeated write cycles by the inductive write head. In the set of heads we tested, the MR head's center and effective width changed only slightly. Although we evaluated the NSIC specification for MR read elements, the specification should be equally valid for other read head types also, as long as the PES patterns are similar     

Automatic MR-PET registration algorithm

We present an algorithm to automatically register magnetic resonance (MR) and positron emission tomographic (PET) images of the human brain. Our algorithm takes an integrated approach: we simultaneously segment the brain in both modalities and register the slices. The algorithm does not attempt to remove the skull from the MR image, but rather uses “templates” constructed from PET images to locate the boundary between the brain and the surrounding tissue in the MR images. The PET templates are a sequence of estimates of the boundary of the brain in the PET images. For each of the templates, the registration algorithm aligns the MR and PET images by minimizing an energy function. The energy function is designed to implicitly model the relevant anatomical structure in the MR image. The template with the lowest energy after registration is the PET brain boundary. The alignment of this template in the MR image marks the MR brain boundary and gives the transformation between the two images. © 1998 John Wiley & Sons, Inc. Int J Imaging Syst Technol, 9, 46–50, 1998     

From ballistic transport to tunneling in electromigrated ferromagnetic breakjunctions

We fabricate ferromagnetic nanowires with constrictions whose cross section can be reduced gradually from 100 x 30 nm(2) to the atomic scale and eventually to the tunneling regime by means of electromigration. The contacts are mechanically and thermally stable. We measure low-temperature magnetoresistances (MR) < 3% for contacts < 400 Omega, reproducible MR variations that are nonmonotonic in the regime 400 Omega - 25 kOmega, and a maximum MR of 80% for atomic-scale widths. These results for devices > 400 Omega differ from previous room-temperature studies of electrodeposited devices. For samples in the tunneling regime, we observe large fluctuations in MR, between -10 and 85%.     

FFCAEs: An efficient feature fusion framework using cascaded autoencoders for the identification of gliomas

Intracranial tumors arise from constituents of the brain and its meninges. Glioblastoma (GBM) is the most common adult primary intracranial neoplasm and is categorized as high-grade astrocytoma according to the World Health Organization (WHO). The survival rate for 5 and 10 years after diagnosis is under 10%, contributing to its grave prognosis. Early detection of GBM enables early intervention, prognostication, and treatment monitoring. Computer-aided diagnostics (CAD) is a computerized process that helps to differentiate between GBM and low-grade gliomas (LGG), using the perceptible analysis of magnetic resonance (MR) of the brain. This study proposes a framework consisting of a feature fusion algorithm with cascaded autoencoders (CAEs), referred to as FFCAEs. Here we utilized two CAEs and extracted the relevant features from multiple CAEs. Inspired by the existing work on fusion algorithms, the obtained features are then fused by using a novel fusion algorithm. Finally, the resultant fused features are classified with the Softmax classifier to arrive at an average classification accuracy of 96.7%, which is 2.45% more than the previously best-performing model. The method is shown to be efficacious thus, it can be useful as a utility program for doctors.     

Estimation of absorption and backscattering coefficients from in situ radiometric measurements: theory and validation in case II waters

A model that relates the coefficients of absorption (a) and backscattering (b(b)) to diffuse attenuation (K(d)), radiance reflectance (R(L)), and the mean cosine for downward irradiance (mu(d)) is presented. Radiance transfer simulations are used to verify the physical validity of the model for a wide range of water column conditions. Analysis of thee radiance transfer simulations suggest that absorption and backscattering can be estimated with average errors of 1% and 3%, respectively, if the value of mu(d) is known with depth. If the input data set is restricted to variables that can be derived from measurements of upward radiance (L(u)) and downward irradiance (E(d)), it is necessary to use approximate values of mu(d). Examination of three different approximation schemes for mu(d) shows that the average error for estimating a and b(b) increases to approximately 13%. We tested the model by using measurements of L(u) and E(d) collected from case II waters off the west coast of Scotland. The resulting estimates of a and b(b) were compared with independent in situ measurements of these parameters. Average errors for the data set were of the order of 10% for both absorption and backscattering.     

Analysis of denoising filters on MRI brain images

The magnetic resonance imaging (MRI) modality is an effective tool in the diagnosis of the brain. These MR images are introduced with noise during acquisition which reduces the image quality and limits the accuracy in diagnosis. Elimination of noise in medical images is an important task in preprocessing and there exist different methods to eliminate noise in medical images. In this article, different denoising algorithms such as nonlocal means, principal component analysis, bilateral, and spatially adaptive nonlocal means (SANLM) filters are studied to eliminate noise in MR. Comparative analysis of these techniques have been with help of various metrics such as signal‐to‐noise ratio, peak signal‐to‐noise ratio (PSNR), mean squared error, root mean squared error, and structure similarity (SSIM). This comparative study shows that the SANLM denoising filter gives the best performance in terms of better PSNR and SSIM in visual interpretation. It also helps in clinical diagnosis of the brain.     

Fast multispectral diffuse optical tomography system for in vivo three-dimensional imaging of seizure dynamics

We describe a multispectral continuous-wave diffuse optical tomography (DOT) system that can be used for in vivo three-dimensional (3-D) imaging of seizure dynamics. Fast 3-D data acquisition is realized through a time multiplexing approach based on a parallel lighting configuration--our system can achieve 0.12 ms per source per wavelength and up to a 14 Hz sampling rate for a full set of data for 3-D DOT image reconstruction. The system is validated using both static and dynamic tissue-like phantoms. An initial in vivo experiment using a rat model of seizure is also demonstrated.     

Hyperspectral remote sensing for shallow waters. 2. Deriving bottom depths and water properties by optimization

In earlier studies of passive remote sensing of shallow-water bathymetry, bottom depths were usually derived by empirical regression. This approach provides rapid data processing, but it requires knowledge of a few true depths for the regression parameters to be determined, and it cannot reveal in-water constituents. In this study a newly developed hyperspectral, remote-sensing reflectance model for shallow water is applied to data from computer simulations and field measurements. In the process, a remote-sensing reflectance spectrum is modeled by a set of values of absorption, backscattering, bottom albedo, and bottom depth; then it is compared with the spectrum from measurements. The difference between the two spectral curves is minimized by adjusting the model values in a predictor-corrector scheme. No information in addition to the measured reflectance is required. When the difference reaches a minimum, or the set of variables is optimized, absorption coefficients and bottom depths along with other properties are derived simultaneously. For computer-simulated data at a wind speed of 5 m/s the retrieval error was 5.3% for depths ranging from 2.0 to 20.0 m and 7.0% for total absorption coefficients at 440 nm ranging from 0.04 to 0.24 m(-1). At a wind speed of 10 m/s the errors were 5.1% for depth and 6.3% for total absorption at 440 nm. For field data with depths ranging from 0.8 to 25.0 m the difference was 10.9% (R(2) = 0.96, N = 37) between inversion-derived and field-measured depth values and just 8.1% (N = 33) for depths greater than 2.0 m. These results suggest that the model and the method used in this study, which do not require in situ calibration measurements, perform very well in retrieving in-water optical properties and bottom depths from above-surface hyperspectral measurements.     

Pure rotational coherent anti-Stokes Raman spectroscopy in mixtures of CO and N2

We present a model for quantitative measurements in binary mixtures of nitrogen and carbon monoxide by the use of dual-broadband rotational coherent anti-Stokes Raman spectroscopy. The model has been compared with experimental rotational coherent anti-Stokes Raman scattering spectra recorded within the temperature range of 294-702 K. Temperatures and concentrations were evaluated by spectral fits using libraries of theoretically calculated spectra. The relative error of the temperature measurements was 1-2%, and the absolute error of the CO concentration measurements was <0.5% for temperatures < or =600 K. For higher temperatures, the gas composition was not chemically stable, and we observed a conversion of CO to CO2. The influence of important spectroscopic parameters such as the anisotropic polarizability and Raman line-broadening coefficients are discussed in terms of concentration measurements. In particular, it is shown that the CO concentration measurement was more accurate if N2-CO and CO-N2 line-broadening coefficients were included in the calculation. The applicability of the model for quantitative flame measurements is demonstrated by measuring CO concentrations in ethylene/air flames.     

Prediction of ethylene content in melt-state random and block polypropylene by near-infrared spectroscopy and chemometrics: comparison of a new calibration transfer method with a slope/bias correction method

This paper reports the prediction of the ethylene content (C2 content) in random polypropylene (RPP) and block polypropylene (BPP) in the melt state by near-infrared (NIR) spectroscopy and chemometrics. NIR spectra of RPP and BPP in the melt states were measured by a Fourier transform near-infrared (FT-NIR) on-line monitoring system. The NIR spectra of RPP and BPP were compared. Partial least-squares (PLS) regression calibration models predicting the ethylene (C2) content that were developed by using each RPP or BPP spectra set separately yielded good results (SECV (standard error of cross validation): RPP, 0.16%; BPP, 0.31%; correlation coefficient: RPP, 0.998; BPP, 0.996). We also built a common PLS calibration model by using both the RPP and the BPP spectra set. The results showed that the common calibration model has larger SECV values than the models based on the RPP or the BPP spectra sets individually and is not practical for the prediction of the C2 content. We further investigated whether a calibration model developed by using the BPP spectra set can predict the C2 contents in the RPP sample set. If this is possible, it can save a significant amount of work and cost. The results showed that the use of the BPP model for the RPP sample set is difficult, and vice versa, because there are some differences in the molar absorption coefficients between the RPP and BPP spectra. To solve this problem, a transfer method from one sample spectra (BPP) set to the other spectra (RPP) set was studied. A difference spectrum between an RPP spectrum and a BPP spectrum was used to transfer from the BPP calibration set to the RPP calibration set. The prediction result (SEP (standard error of prediction), 0.23%, correlation coefficient, 0.994) of RPP samples by the transferred calibration set and model showed that it is possible to transfer from the BPP calibration set to the RPP calibration set. We also studied the transfer from the RPP calibration set (the range of C2 content: 0-4.3%) to the BPP calibration set. The prediction result of C2 content (the range of C2 contents: 0-7.7%) in BPP by use of the calibration model based on the transferred BPP spectra from the RPP spectra showed that the transfer method is only effective for the interpolation of the C2 content range by the nonlinear change in the peak intensities with the C2 content.     

An Improved Jellyfish Algorithm for Multilevel Thresholding of Magnetic Resonance Brain Image Segmentations

Image segmentation is vital when analyzing medical images, especially magnetic resonance (MR) images of the brain. Recently, several image segmentation techniques based on multilevel thresholding have been proposed for medical image segmentation; however, the algorithms become trapped in local minima and have low convergence speeds, particularly as the number of threshold levels increases. Consequently, in this paper, we develop a new multilevel thresholding image segmentation technique based on the jellyfish search algorithm (JSA) (an optimizer). We modify the JSA to prevent descents into local minima, and we accelerate convergence toward optimal solutions. The improvement is achieved by applying two novel strategies: Ranking-based updating and an adaptive method. Ranking-based updating is used to replace undesirable solutions with other solutions generated by a novel updating scheme that improves the qualities of the removed solutions. We develop a new adaptive strategy to exploit the ability of the JSA to find a best-so-far solution; we allow a small amount of exploration to avoid descents into local minima. The two strategies are integrated with the JSA to produce an improved JSA (IJSA) that optimally thresholds brain MR images. To compare the performances of the IJSA and JSA, seven brain MR images were segmented at threshold levels of 3, 4, 5, 6, 7, 8, 10, 15, 20, 25, and 30. IJSA was compared with several other recent image segmentation algorithms, including the improved and standard marine predator algorithms, the modified salp and standard salp swarm algorithms, the equilibrium optimizer, and the standard JSA in terms of fitness, the Structured Similarity Index Metric (SSIM), the peak signal-to-noise ratio (PSNR), the standard deviation (SD), and the Features Similarity Index Metric (FSIM). The experimental outcomes and the Wilcoxon rank-sum test demonstrate the superiority of the proposed algorithm in terms of the FSIM, the PSNR, the objective values, and the SD; in terms of the SSIM, IJSA was competitive with the others.