Normalization of brain magnetic resonance images using histogram even-order derivative analysis

The even-ordered (2nd, 4th and 6th) derivatives of a brain MRI histogram were used to calculate a characteristic value for white matter, which was used to normalize the image intensity scale. Simulated image histograms were used to estimate the methodological error as a function of noise level, and the optimum derivative order was determined for each image type studied (T1-, T2- and density-weighted). The algorithm yielded highly reproducible results when used in conjunction with a threshold-sensitive brain segmentation algorithm. It also proved insensitive to the presence of extra-cranial tissues. This method of histogram analysis could find utility in a variety of applications that demand robust intensity normalization including image registration, brain segmentation, tissue classification and spatial inhomogeneity correction.     

A comparison study of inhomogeneous magnetization transfer (ihMT) and magnetization transfer (MT) in multiple sclerosis based on whole brain acquisition at 3.0 T

IntroductionMultiple sclerosis (MS) is a central nervous system disorder that may eventually affect its function. The clinical standard for MS severity is based on a clinical scale, which lacks lesion specific information. Magnetic resonance imaging of MS faces the challenge of myelin specificity, and in this work a new method inhomogeneous magnetization transfer (ihMT) is investigated as new biomarker of demyelination in MS.MethodsLocal ethics committee approved this study and written informed consents were obtained. Between Oct 2017 to May 2018, eighteen patients with relapsing-remitting MS (RRMS) (6 males, 12 females, mean age 31.2) and sixteen healthy volunteers (6 males, 10 females, mean age 30.4 years) were enrolled in this prospective study. All subjects underwent MRI exams including MT and ihMT imaging as well as the Expanded Disability Status Scale (EDSS) assessments. Independent sample t-test were used to compare the difference of ihMT parameters between healthy white matter (HWM) and normal appearing white matter (NAWM) and between HWM and MS lesions, respectively. Spearman correlation were used to analyze the correlation between ihMT parameters of MS lesions and EDSS score.ResultsThe ihMTR and qihMT demonstrate significant differences between WHM and NAWM groups, while no significant differences are observed for MTR and qMT. All parameters show significant differences between HWM and MS groups (p < 0.05). There was moderate negative correlation between MTR, qMT and EDSS score (−0.440 and −0.572), while there was a strong negative correlation between ihMTR and qihMT and EDSS score (−0.704 and −0.739).ConclusionBased on whole brain analysis at 3.0 T, ihMT showed better correlation with EDSS compared to magnetization transfer imaging, and may be a potentially valuable biomarker for demyelination in MS.     

Combined volumetric T1, T2 and secular-T2 quantitative MRI of the brain: age-related global changes (preliminary results)

The combined T1, T2 and secular-T2 pixel frequency distributions of 24 adult human brains were studied in vivo using a technique based on the mixed-TSE pulse sequence, dual-space clustering segmentation and histogram gaussian decomposition. Pixel frequency histograms of whole brains and the four principal brain compartments were studied comparatively and as function of age. For white matter, the position of the T1 peak correlates with age (R2 =.7868) when data are fitted to a quadratic polynomial. For gray matter, a weaker age correlation is found (R2 =.3687). T2 and secular-T2 results are indicative of a weaker correlation with age. The technique and preliminary results presented herein may be useful for characterizing normal as well as abnormal aging of the brain, and also for comparison with the results obtained with alternative quantitative MRI methodologies.     

Detection of Apple Marssonina Blotch with PLSR,PCA, and LDA Using Outdoor Hyperspectral Imaging

In this study, hyperspectral images were used to detect a fungal disease in apple leaves called Marssonina blotch(AMB). Estimation models were built to classify healthy, asymptomatic and symptomatic classes using partial least squares regression(PLSR), principal component analysis(PCA), and linear discriminant analysis(LDA) multivariate methods. In general, the LDA estimation model performed the best among the three models in detecting AMB asymptomatic pixels, while all the models were able to detect the symptomatic class. LDA correctly classified asymptomatic pixels and LDA model predicted them with an accuracy of 88.0%. An accuracy of 91.4% was achieved as the total classification accuracy. The results from this work indicate the potential of using the LDA estimation model to identify asymptomatic pixels on leaves infected by AMB.     

Different classification algorithms and serum surface enhanced Raman spectroscopy for noninvasive discrimination of gastric diseases

In this study, surface enhanced Raman spectroscopy (SERS) was used to investigate the spectral characteristics of blood serum for the purpose of diagnosing stomach diseases. SERS spectral data was collected from patients with atrophic gastritis, both pre‐operation and post‐operation gastric cancer, and from healthy individuals. Visual differences in the SERS spectra were observed between the four groups which indicate corresponding biomolecule concentration changes in blood. To further investigate the diagnostic ability of human serum, the spectral data was analyzed with three chemometric processes. These three methods extracted features and classified from the spectral data. Principal component analysis (PCA) was first performed to reduce the dimensionality of the original spectral data. Then, the classification methods support vector machine (SVM), linear discriminant analysis (LDA) and classification and regression tree (CART) were used for the evaluation of diagnostic ability. Accuracies of 96.5%, 88.8% and 87.1% were obtained for PCA‐SVM, PCA‐LDA and PCA‐CART, respectively. Copyright © 2016 John Wiley & Sons, Ltd.     

近红外光谱指纹分析在羊肉产地溯源中的应用

为寻求低廉、快速有效地签别羊肉产地来源的方法,对来自内蒙古自治区锡林郭勒盟、呼伦贝尔市和阿拉善盟三个牧区,及重庆市和山东省菏泽市两个农区共99份羊肉样品进行近红外光谱扣描,利用主成分分析结合线性判别分析(PCA+LDA),以及偏最小二乘判别分析法(PLS-DA)对光谱数据进行了分析,建立了羊肉产地来源的定性判别模型.结...     

Magnetic resonance of brain tumors: considerations of imaging contrast on the basis of relaxation measurements

Proton spin-lattice and spin-spin relaxation times have been measured in surgically-removed normal CNS tissues and a variety of tumors of the brain. All measurements were made at 20 MHz and 37 degrees C. Between grey and white matter from autopsy human or canine specimens significant differences in T1 or T2 were observed, with greater differences seen in T1. Such discrimination was reduced in samples obtained from live brain-tumor patients due to lengthening in T1 and T2 of white matter near tumorous lesions. Edematous white matter showed T1 and T2 values higher than those of autopsy disease-free white matter. Compared to normal CNS tissues, most brain tumors examined in this study demonstrated elevated T1 and T2 values. Exceptions, however, did exist. No definitive correlation was indicated on a T1 or T2 basis which allowed a distinction to be made between benign and malignant states. Furthermore, considerable variation in relaxation times occurred from tumor to tumor of the same type, suggesting that within a tumor type there are important differences in physiology, biology, and/or pathologic state. Such variation caused partial overlap in relaxation times among certain tumor types and hence may limit the capability of magnetic resonance imaging (MR) alone for the diagnosis of specific disease. Nonetheless, this study predicts that on the basis of T1 or T2 differences most brain tumors are readily detectable by MR via saturation recovery or inversion recovery with appropriate selections of pulse-spacing parameters. In general, tumors can be discriminated against white matter better than grey matter and contrast between glioma and grey matter is usually superior to that between meningioma and grey matter. This work did not consider tissue-associated proton density which should be addressed together with T1 and T2 for a complete treatment of MR contrast.     

高光谱成像的机采籽棉杂质分类检测

机采籽棉杂质分类检测为调整棉花清理机械加工参数和工序提供参考依据,对提升皮棉品质具有重要意义。但由于籽棉棉层分布不均匀,使得图像检测难度增大,使用传统的检测方法无法有效检测各类杂质。采用高光谱成像方法对机采籽棉中的棉叶、棉枝、地膜和铃壳（内外）五种杂质进行分类判别检测。首先采集120个机采籽棉样本的高光谱图像,选取感兴趣区域获取平均光谱曲线。发现由于物质构成的差异,不同杂质体现出不同的吸收和反射特性,不同种类物质之间的光谱差异大于同类物质。对提取的平均光谱曲线进行主成分分析（PCA）,结果显示棉花、残膜和铃壳外与其他三类相比,有较好的聚集性和可分性,但是棉叶、铃壳内和棉枝三类相互叠加在一起,空间分布存在严重交叉重叠。以提取的平均光谱曲线为训练样本,选择线性判别分析（LDA）、支持向量机（SVM）和神经网络（ANN）三种分类判别算法,对算法参数进行寻优,并建立机采籽棉杂质分类判别模型。其中,经过LDA模型降维后的样本空间较PCA表现出了更好的聚集性和可分性,采用正则化防止过拟合,得到训练集准确率为86.4%,测试集准确率为86.2%;SVM模型的参数寻优结果为C=105,g=0.1,其训练集准确率为83.42%,测试集准确率为83.40%;ANN模型参数寻优得到隐含层数和神经元个数分别为2和17,训练集准确率为82.9%,测试集准确率为81.8%。对三种模型的分类效果和检测用时进行比较,LDA模型结果最优。通过对高光谱图像进行像素等级分类判别,结果显示棉花识别效果较好,植物性杂质都被有效检测,但是地膜和棉花存在误识别,分类效果与杂质光谱的分类判别模型结果一致。因此,采用高光谱成像技术可以快速、无损的检测和识别籽棉杂质,为棉花加工装备提供反馈参数,对棉花加工机械化和智能化有重要意义。     

Regional variation in rat brain proton relaxation times and water content

Relaxation times (T1 and T2) and water content are measured in frontal cortex, amygdaloid cortex, hippocampus, mid-brain and cerebellum of rat brain. Differences are found in relaxation times, between areas containing a mixture of grey and white matter, and grey matter only. Differences were also found between certain grey matter areas. Relaxation times correlated with water content.     

Morphometry in temporal lobe epilepsy

We demonstrate a method for quantitating changes in volume and morphology of the temporal lobe in epilepsy. The temporal lobes of 10 neurologically normal subjects and six subjects with well defined left-sided mesial temporal epilepsy were studied. From high resolution T₁-weighted magnetic resonance images, the grey and white matter were manually segmented over a predetermined extent. The volumes of the grey and white matter were determined. Using the segmented images, the grey matter/CSF surface and the white matter/grey matter surface were reconstructed, allowing estimates of the surface area and calculation of indices of curvature for the two surfaces. The index of curvature was calculated for each vertex of a polygonal mesh that was fitted to the surfaces. An index of grey matter thickness (grey matter volume/white matter surface area) was also calculated. There was a significant bilateral decrease in the total volume (p < .01), grey matter volume (p < .001) and grey matter thickness index (p < .05) in epileptic subjects. In addition, there was a bilateral decrease in white matter surface area (p < .05) and a small left-sided decrease in white matter volume (p < .05) in epileptic subjects. The average distributions of indices of curvature for both surfaces differed significantly (p < .05) between normal and epileptic subjects. In the grey matter/CSF surface of normal subjects, a large peak corresponding to surface concavity was present. The amplitude of this peak was significantly lower in epileptic subjects (p < .05 for the right hemisphere; p < .001 for the left hemisphere).     

Reduced field-of-view DTI segmentation of cervical spine tissue

The number of diffusion tensor imaging (DTI) studies regarding the human spine has considerably increased and it is challenging because of the spine’s small size and artifacts associated with the most commonly used clinical imaging method. A novel segmentation method based on the reduced field-of-view (rFOV) DTI dataset is presented in cervical spinal canal cerebrospinal fluid, spinal cord grey matter and white matter classification in both healthy volunteers and patients with neuromyelitis optica (NMO) and multiple sclerosis (MS). Due to each channel based on high resolution rFOV DTI images providing complementary information on spinal tissue segmentation, we want to choose a different contribution map from multiple channel images. Via principal component analysis (PCA) and a hybrid diffusion filter with a continuous switch applied on fourteen channel features, eigen maps can be obtained and used for tissue segmentation based on the Bayesian discrimination method. Relative to segmentation by a pair of expert readers, all of the automated segmentation results in the experiment fall in the good segmentation area and performed well, giving an average segmentation accuracy of about 0.852 for cervical spinal cord grey matter in terms of volume overlap. Furthermore, this has important applications in defining more accurate human spinal cord tissue maps when fusing structural data with diffusion data. rFOV DTI and the proposed automatic segmentation outperform traditional manual segmentation methods in classifying MR cervical spinal images and might be potentially helpful for detecting cervical spine diseases in NMO and MS.     

Relaxo-volumetric multispectral quantitative magnetic resonance imaging of the brain over the human lifespan: global and regional aging patterns

The objective of this study was to determine the T1, T2 and secular-T2 relaxo-volumetric brain aging patterns using multispectral quantitative magnetic resonance imaging, both globally and regionally, and covering an age range approaching the full human lifespan. Fifty-one subjects (28 males, 23 females; age range: 0.5–87 years) were studied consisting of 18 healthy volunteers and 33 patients. Patients were selected after carefully reviewing their radiology reports to have either normal-by-MRI findings (25 patient subjects) or small focal pathology less than 6 mm in size (eight patient subjects). All subjects were MR imaged at 1.5 T with the mixed turbo spin echo pulse sequence. The soft tissues inside the cranial vault, termed intracranial matter (ICM), were segmented using a dual-clustering segmentation algorithm. ICM segments were further divided into six subsegments: bilateral anterior cerebral, posterior cerebral and cerebellar subsegments. T1, T2 and secular-T2 relaxation time histograms of all segments were generated and modeled with Gaussian functions. For each segment, the volumes of white matter, gray matter and cerebrospinal fluid were calculated from the T1 histograms. The age-related tendencies of three quantitative MRI parameters (T1, T2 and secular-T2) and the fractional tissue volumes showed four distinct periods of life, specifically a maturation period (0–2 years), a development period (2–20 years), an adulthood period (20–60 years) and a senescence period (60 years and older). For all ages, the anterior cerebral subsegment exhibited consistently longer gray matter T1s and shorter white matter T1s than the posterior cerebral and cerebellar subsegments. Volumetric age-related changes of the cerebellar subsegment were more gradual than in the cerebral subsegments. This study shows that relaxometric and volumetric age-related changes are synchronized and define the same four periods of brain evolution both globally and regionally.     

Ex vivo magnetic resonance imaging of rat spinal cord at 9.4 T

The magnetic resonance (MR) properties of the rat spinal cord were characterized at the T9 level with ex vivo experiments performed at 9.4 T. The inherent endogenous contrast parameters, proton density (PD), longitudinal and transverse relaxation times T1 and T2, and magnetization transfer ratio (MTR) were measured separately for the grey matter (GM) and white matter (WM). Analysis of the measurements indicated that these tissues have statistically different proton densities with means PD(GM)=54.8+/-2.5% versus PD(WM)=45.2+/-2.4%, and different T1 values with means T1GM=2.28+/-0.23 s versus T1WM=1.97+/-0.21 s. The corresponding values for T2 were T2GM=31.8+/-4.9 ms versus T2WM=29.5+/-4.9 ms, and the difference was insignificant. The difference between MTR(GM)=31.2+/-6.1% and MTR(WM)=33.1+/-5.9% was also insignificant. These results collectively suggest that PD and T1 are the two most important parameters that determine the observed contrast on spinal cord images acquired at 9.4 T. Therefore, in MR imaging studies of spinal cord at this field strength, these parameters need to be considered not only in optimizing the protocols but also in signal enhancement strategies involving exogenous contrast agents.     

T2* measurements in human brain at 1.5, 3 and 7 T

Measurements have been carried out in six subjects at magnetic fields of 1.5, 3 and 7 T, with the aim of characterizing the variation of T2* with field strength in human brain. Accurate measurement of T2* in the presence of macroscopic magnetic field inhomogeneity is problematic due to signal decay resulting from through-slice dephasing. The approach employed here allowed the signal decay due to through-slice dephasing to be characterized and removed from data, thus facilitating an accurate measurement of T2* even at ultrahigh field. Using double inversion recovery turbo spin-echo images for tissue classification, an analysis of T2* relaxation times in cortical grey matter and white matter was carried out, along with an evaluation of the variation of T2* with field strength in the caudate nucleus and putamen. The results show an approximately linear increase in relaxation rate R2* with field strength for all tissues, leading to a greater range of relaxation times across tissue types at 7 T that can be exploited in high-resolution T2*-weighted imaging.     

基于时频谱特征的白酒品质分类方法研究

为了建立快速、准确的白酒品质鉴别方法,利用机器学习方法对不同品质的白酒建模。为了提取不同品质白酒的特征,使用离子迁移谱对不同品质白酒进行分析,构建了基于白酒离子迁移谱信号的特征向量,并对不同品质的白酒进行了识别与分类。白酒样本的离子迁移谱信号通过利用美国Excellims公司GA2100型电喷雾-离子迁移谱仪（ESI-IMS）采集获得,每一个离子迁移谱信号是强度随时间变化的时间序列信号;提取了原始数据离子迁移谱的时域特征谱峰。为了获得更全面的特征,对离子迁移谱数据进行了傅里叶变换并提取频域内的特征谱峰。同时为了表述信号变化的特征,计算了离子迁移谱的谱熵和过零率,构建N×9维的特征向量矩阵;使用主成分分析(PCA)和线性判别分析(LDA)分别对上述获得的特征进行了特征降维,其中使用PCA对特征向量矩阵降维后的前三维特征对整体特征的累计贡献率达到了95%,而使用LDA对特征向量矩阵降维后的前两维特征对整体特征的累计贡献率就达到了95%。因此,选择了LDA作为特征降维方法;最后,利用机器学习中的非线性分类器支持向量机(SVM)对白酒离子迁移谱数据进行分类研究。实验结果表明,在真酒和添加酒精的白酒二分类中,SVM方法正确分类率达到100%;而在真酒和分别添加10%,20%,30%,40%和50%酒精浓度的五种假酒的六分类中,SVM方法正确分类率达到99.7%。比较了逻辑回归(LRM)分类、模糊C均值分类(FCM)和K近邻分类(KNN)对白酒样本离子迁移谱分类实验结果。研究表明,对于离子迁移谱非常接近的真酒和添加酒精的白酒,基于频谱特征向量的SVM方法能够准确的区分开来,为白酒的品质鉴别提供了一种新的检测方法。     

基于显微拉曼对氮胁迫下微藻油脂变化的研究

利用共聚焦显微拉曼光谱仪获取生长在三种氮营养条件下(氮胁迫、氮正常、氮饱和)培养的蛋白核小球藻(Chlorella pyrenoidosa)的拉曼光谱,通过拉曼散射光谱信息对微藻在不同氮胁迫下生长情况及油脂变化进行研究。对油脂拉曼特征峰值比值作气泡图以直观表达油脂积累量,该气泡图与尼罗红荧光图像具有良好的相关性。光谱信号经预处理后,利用主成分分析(PCA)对全波段进行分析,获得相应的主成分变量,通过线性判别分析(LDA)建立分类模型。利用PCA获取的主成分变量建立的LDA预测模型对三种氮营养条件的预测正确率分别是80%, 93.3%, 86.7%。基于油脂特征位移(RS)处的比率建立的LDA分类模型对三种氮营养条件的分类正确率最高达到86.7%。研究结果表明,利用拉曼技术对微藻生长的不同氮胁迫条件鉴别是可行的,且随着氮胁迫影响的时间增加,油脂的积累差异就越大。     

基于广义判别分析的光谱分类

提出了基于广义判别分析(generalized discriminant analysis, GDA)方法对恒星(Star)、星系(Galaxy)和类星体(Quasars)的光谱进行分类。广义判别分析将核技巧与Fisher判别分析结合起来,通过非线性映射将样本集映射到高维特征空间F,在F空间中进行线性判别分析。实验对比了LDA, GDA, PCA, KPCA算法对于恒星、星系和类星体的光谱分类性能。结果表明基于GDA的算法对于这3种类型光谱的分类正确率最高,LDA次之;尽管KPCA也是一种基于核的方法,但是选择主成分个数较少时效果较差,甚至低于LDA;基于PCA的分类效果最差。     

High resolution neuroimaging at 4.1 T

In this article we report on acquisition of high resolution 512 × 512 images at 4.1 T using an inversion recovery gradient-echo sequence and a volume head coil developed for high field applications. The T₁ values for cerebral white and grey matter were measured to be 834 and 1282 ms, respectively. The partial saturation inversion recovery sequence (T_ir 800 ms and TR 2500 ms) provided excellent contrast-to-noise for white to grey matter. Consequently, the images consistently visualized the thalamic nuclear groups, hippocampal fine structure, as well as small draining vessels of the white matter.     

Segmenting brain white matter, gray matter and cerebro-spinal fluid using diffusion tensor-MRI derived indices.

Based on its ability to provide quantitative information about tissue microstructure, diffusion tensor magnetic resonance imaging (DT-MRI) might be a valuable approach to improve the reliability of segmentation of the various brain tissues.In this study, a fully automated and easy-to-implement technique based on 2D histogram analysis of DT-MRI derived images was used to segment white and gray matter of the brain from 10 healthy subjects (aged = 27-56 years). The results obtained with this novel segmentation strategy were compared to those achieved by two experienced observers using an operator-dependent segmentation on the dual-echo scans.Visual inspection of the segmented tissues from a third senior observer disclosed that the automated technique worked properly on all images from all subjects and was more accurate than the human raters in defining thalamus white and gray matter portions as well as in tissue classification at the external brain edge. In addition, this segmentation technique resulted in an average gray/white matter ratio similar to that reported by post-mortem assessment. The application of the operator-dependent segmentation strategy was extremely time-consuming and the two observers achieved poorly reproducible results.Segmenting brain white and gray matter using information from DT-MRI proved to be an accurate approach with the potential for improving the understanding of the pathophysiology of many neurologic conditions.     

Estimation of the content of fat and parenchyma in breast tissue using MRI T1 histograms and phantoms

Mammographic breast density has been correlated with breast cancer risk. Estimation of the volumetric composition of breast tissue using three-dimensional MRI has been proposed, but accuracy depends upon the estimation methods employed. The use of segmentation based on T1 relaxation rates allows quantitative estimates of fat and parenchyma volume, but is limited by partial volume effects. An investigation employing phantom breast tissue composed of various combinations of chicken breast (to represent parenchyma) and cooking fats was carried out to elucidate the factors that influence MRI T1 histograms. Using the phantoms, T1 histograms and their known fat and parenchyma composition, a logistic distribution function was derived to describe the apportioning of the T1 histogram to fat and parenchyma. This function and T1 histograms were then used to predict the fat and parenchyma content of breasts from 14 women. Using this method, the composition of the breast tissue in the study population was as follows: fat 69.9+/-22.9% and parenchyma 30.1+/-22.9%.