First-order Statistics Research Articles

Artificial intelligence-based myocardial texture analysis in etiological differentiation of left ventricular hypertrophy.

BackgroundTransthoracic echocardiography (TTE) is widely used in clinics to evaluate left ventricular hypertrophy (LVH). However, TTE is usually insufficient for the etiological diagnoses when morphological and functional features are nonspecific. With the booming of computer science and artificial intelligence (AI), previous literature has reported the application of radiomics based on cardiac magnetic resonance imaging, cardiac computed tomography and TTE in diagnosing several myocardial abnormalities, such as myocardial infarction, myocarditis, cardiac amyloidosis, and hypertrophic cardiomyopathy (HCM). In this study, we explored the possibility of using myocardial texture features in differentiating HCM, hypertensive heart disease (HHD) and uremic cardiomyopathy (UCM) based on echocardiography. To our knowledge, this was the first study to explore TTE myocardial texture analysis for multiple LVH etiology differentiation.MethodsTTE images were reviewed retrospectively from January 2018 to collect 50 cases for each group of HHD, HCM and UCM. The apical four chamber view was retrieved. Seventeen first-order statistics and 60 gray level co-occurrence matrix (GLCM) features were extracted for statistics and classification test by support vector machine (SVM).ResultsOf all the parameters, entropy of brightness (EtBrt), standard deviation (Std), coefficient of variation (CoV), skewness (Skew), contrast7 (Cont7) and homogeneity5 (Hm5) were found statistically significant among the three groups (all P<0.05) and with acceptable reproducibility (intraobserver and interobserver ICC >0.50). As a result, HCM showed the most homogeneous myocardial texture, and was significantly different from HHD and UCM (all six features: P≤0.005). HHD appeared slightly more homogeneous than UCM, as only EtBrt and CoV were significant (P=0.011 and P=0.008). According to higher areas under the receiver operating characteristic curve (AUC) (>0.50), EtBrt, Std, and CoV were selected for test of classification as a combination of features. The AUC derived from SVM model was slightly improved compared with those of EtBrt, Std and CoV individually.ConclusionsAI-based myocardial texture analysis using ultrasonic images may be a potential approach to aiding LVH etiology differentiation.

Residual Augmented Attentional U-Shaped Network for Spectral Reconstruction from RGB Images

Deep convolutional neural networks (CNNs) have been successfully applied to spectral reconstruction (SR) and acquired superior performance. Nevertheless, the existing CNN-based SR approaches integrate hierarchical features from different layers indiscriminately, lacking an investigation of the relationships of intermediate feature maps, which limits the learning power of CNNs. To tackle this problem, we propose a deep residual augmented attentional u-shape network (RA2UN) with several double improved residual blocks (DIRB) instead of paired plain convolutional units. Specifically, a trainable spatial augmented attention (SAA) module is developed to bridge the encoder and decoder to emphasize the features in the informative regions. Furthermore, we present a novel channel augmented attention (CAA) module embedded in the DIRB to rescale adaptively and enhance residual learning by using first-order and second-order statistics for stronger feature representations. Finally, a boundary-aware constraint is employed to focus on the salient edge information and recover more accurate high-frequency details. Experimental results on four benchmark datasets demonstrate that the proposed RA2UN network outperforms the state-of-the-art SR methods under quantitative measurements and perceptual comparison.

Feature-based automated segmentation of ablation zones by fuzzy c-mean clustering during low-dose computed tomography.

Intra-procedural monitoring and post-procedural follow-up is necessary for a successful ablation treatment. An imaging technique which can assess the ablation geometry accurately is beneficial to monitor and evaluate treatment. In this study, we developed an automated ablation segmentation technique for serial low-dose, noisy ablation computed tomography (CT) or contrast-enhanced CT (CECT). Low-dose, noisy temporal CT and CECT volumes were acquired during microwave ablation on normal porcine liver (four with non-contrast CT and eight with CECT). Highly constrained backprojection (HYPR) processing was used to recover ablation zone information compromised by low-dose noise. First-order statistic features and normalized fractional Brownian features (NBF) were used to segment ablation zones by fuzzy c-mean clustering. After clustering, the segmented ablation zone was refined by cyclic morphological processing. Automatic and manual segmentations were compared to gross pathology with Dice's coefficient (morphological similarity), while cross-sectional dimensions were compared by percent difference. Automatic and manual segmentations of the ablation zone were very similar to gross pathology (Dice Coefficients: Auto.-Path.=0.84±0.02; Manu.-Path.=0.76±0.03, P=0.11). The differences in ablation area, major diameter and minor diameter were 17.9±3.2%, 11.1±3.2% and 16.2±3.4%, respectively, when comparing automatic segmentation to gross pathology, which were lower than the differences of 32.9±16.8%, 13.0±9.8% and 21.8±5.8% when comparing manual segmentation to gross pathology. Manual segmentations tended to overestimate gross pathology when ablation area was less than 15 cm2 , but the automated segmentation tended to underestimate gross pathology when ablation zone is larger than 20 cm2 . Fuzzy c-means clustering may be used to aid automatic segmentation of ablation zones without prior information or user input, making serial CT/CECT has more potential to assess treatments intra-procedurally.

Optimized recurrent neural network with fuzzy classifier for data prediction using hybrid optimization algorithm: Scope towards diverse applications

Today, data processing has become a challenging task due to the significant increase in the amount of data collected using various sensors. To put up knowledge and forecast the data, the existing data mining techniques compute all numerical attributes in the memory simultaneously. However, the over-abundance of entire factors in the data makes accurate prediction infeasible. This paper attempts to implement a new data prediction model using an optimized machine learning algorithm. The proposed data prediction model involves four main phases: (a) data acquisition, (b) feature extraction, (c) data normalization, and (d) prediction. Initially, few data from the UCI repository like Bike Sharing Dataset, Carbon Nanotubes, Concrete Compressive Strength, Electrical Grid Stability Simulated Data, and SkillCraft-1 Master Table are collected. Further, the feature extraction process extracts the first-order statistics like mean, median, standard deviation, the maximum value of entire data, and the minimum value of entire data, and the second-order statistics like kurtosis, skewness, energy, and entropy. Next, the data or feature normalization is done to arrange the data within a certain limit. The normalized features are then subjected to a hybrid prediction system by integrating the Recurrent Neural Network (RNN) and Fuzzy Regression model. As a modification, the number of hidden neurons in the RNN and membership limits of the Fuzzy Regression model are optimized by a hybrid optimization algorithm by merging the concepts of Whale Optimization Algorithm (WOA) and Cat Swarm Optimization (CSO), which is called the Whale Updated Seek Mode-based CSO (WS-CSO) algorithm. Then, the efficiency of the optimized hybrid classifier for all-time prediction of data in different applications is confirmed based on its valuable performance and comparative analysis.

IMG-13. MRI-BASED RADIOMICS PROGNOSTIC MARKERS OF POSTERIOR FOSSA EPENDYMOMA

PURPOSEPosterior fossa ependymomas (PFE) are common pediatric brain tumors often assessed with MRI before surgery. Advanced radiomic analysis show promise in stratifying risk and outcome in other pediatric brain tumors. Here, we extracted high-dimensional MRI features to identify prognostic, image-based, radiomics markers of PFE and compared its performance to clinical variables.METHODS93 children from five centers (median age=3.3yrs; 59 males; mean PFS=50mos) were included. Tumor volumes were manually contoured on T1-post contrast and T2-weighted MRI for PyRadiomics feature extraction. Features include first-order statistics, size, shape, and texture metrics calculated on the original, log-sigma, and wavelet transformed images. Progression free survival (PFS) served as outcome. 10-fold cross-validation of a LASSO Cox regression was used to predict PFS. Model performance was analyzed and concordance metric (C) was determined using clinical variable (age at diagnosis and sex) only, radiomics only, and radiomics plus clinical variable.RESULTSSix radiomic features were selected (all T1): 1 first-order kurtosis (log-sigma) and 5 texture features (3 wavelet, 2 original). This model demonstrated significantly higher performance than a clinical model alone (C: 0.69 vs 0.58, p<0.001). Adding clinical features to the radiomic features didn’t improve prediction (p=0.67). For patients with molecular subtyping (n=48), adding this feature to the clinical plus radiomics models significantly improved performance over clinical features alone (C = 0.79 vs. 0.66, p=0.02). Further validation and model refinement with additional datasets are ongoing.CONCLUSIONOur pilot study shows potential role for MRI-based radiomics and machine learning for PFE risk stratification and as radiographic biomarkers.

Combination of temporal‐channels correlation information and bilinear feature for action recognition

In this study, the authors focus on improving the spatio–temporal representation ability of three‐dimensional (3D) convolutional neural networks (CNNs) in the video domain. They observe two unfavourable issues: (i) the convolutional filters only dedicate to learning local representation along input channels. Also they treat channel‐wise features equally, without emphasising the important features; (ii) traditional global average pooling layer only captures first‐order statistics, ignoring finer detail features useful for classification. To mitigate these problems, they proposed two modules to boost 3D CNNs’ performance, which are temporal‐channel correlation (TCC) and bilinear pooling module. The TCC module can capture the information of inter‐channel correlations over the temporal domain. Moreover, the TCC module generates channel‐wise dependencies, which can adaptively re‐weight the channel‐wise features. Therefore, the network can focus on learning important features. With regards to the bilinear pooling module, it can capture more complex second‐order statistics in deep features and generate a second‐order classification vector. We can get more accurate classification results by combining the first‐order and second‐order classification vector. Extensive experiments show that adding our proposed modules to I3D network could consistently improve the performance and outperform the state‐of‐the‐art methods. The code and models are available at https://github.com/caijh33/I3D_TCC_Bilinear.

Toward Measuring Target Perception: First-Order and Second-Order Deep Network Pipeline for Classification of Fixation-Related Potentials

The topdown determined visual object perception refers to the ability of a person to identify a prespecified visual target. This paper studies the technical foundation for measuring the target-perceptual ability in a guided visual search task, using the EEG-based brain imaging technique. Specifically, it focuses on the feature representation learning problem for single-trial classification of fixation-related potentials (FRPs). The existing methods either capture only first-order statistics while ignoring second-order statistics in data, or directly extract second-order statistics with covariance matrices estimated with raw FRPs that suffer from low signal-to-noise ratio. In this paper, we propose a new representation learning pipeline involving a low-level convolution subnetwork followed by a high-level Riemannian manifold subnetwork, with a novel midlevel pooling layer bridging them. In this way, the discriminative power of the first-order features can be increased by the convolution subnetwork, while the second-order information in the convolutional features could further be deeply learned with the subsequent Riemannian subnetwork. In particular, the temporal ordering of FRPs is well preserved for the components in our pipeline, which is considered to be a valuable source of discriminant information. The experimental results show that proposed approach leads to improved classification performance and robustness to lack of data over the state-of-the-art ones, thus making it appealing for practical applications in measuring the target-perceptual ability of cognitively impaired patients with the FRP technique.

Makine Öğrenimi Tekniklerini Kullanarak SAR Görüntülemesinden Otomatik Hedef Tanıma (OHT)

Automatic Target Recognition (ATR) in Synthetic aperture radar (SAR) images becomes a very challenging problem owing to containing high level noise. In this study, a machine learning-based method is proposed to detect different moving and stationary targets using SAR images. First Order Statistical (FOS) features were obtained from Fast Fourier Transform (FFT), Discrete Cosine Transform (DCT) and Discrete Wavelet Transform (DWT) on gray level SAR images. Gray Level Co-occurrence Matrix (GLCM), Gray Level Run Length Matrix (GLRLM) and Gray Level Size Zone Matrix (GLSZM) algorithms are also used. These features are provided as input for the training and testing stage Support Vector Machine (SVM) model with Gaussian kernels. 4-fold cross-validations were implemented in performance evaluation. Obtained results showed that GLCM + SVM algorithm is the best model with 95.26% accuracy. This proposed method shows that moving and stationary targets in MSTAR database could be recognized with high performance.

A comprehensive study of brain tumour discrimination using phase combinations, feature rankings, and hybridised classifiers.

The binary categorisation of brain tumours is challenging owing to the complexities of tumours. These challenges arise because of the diversities between shape, size, and intensity features for identical types of tumours. Accordingly, framework designs should be optimised for two phenomena: feature analyses and classification. Based on the challenges and difficulty of the issue, limited information or studies exist that consider the binary classification of three-dimensional (3D) brain tumours. In this paper, the discrimination of high-grade glioma (HGG) and low-grade glioma (LGG) is accomplished by designing various frameworks based on 3D magnetic resonance imaging (3D MRI) data. Accordingly, diverse phase combinations, feature-ranking approaches, and hybrid classifiers are integrated. Feature analyses are performed to achieve remarkable performance using first-order statistics (FOS) by examining different phase combinations near the usage of single phases (T1c, FLAIR, T1, and T2) and by considering five feature-ranking approaches (Bhattacharyya, Entropy, Roc, t test, and Wilcoxon) to detect the appropriate input to the classifier. Hybrid classifiers based on neural networks (NN) are considered due to their robustness and superiority with medical pattern classification. In this study, state-of-the-art optimisation methods are used to form the hybrid classifiers: dynamic weight particle swarm optimisation (DW-PSO), chaotic dynamic weight particle swarm optimisation (CDW-PSO), and Gauss-map-based chaotic particle-swarm optimisation (GM-CPSO). The integrated frameworks, including DW-PSO-NN, CDW-PSO-NN, and GM-CPSO-NN, are evaluated on the BraTS 2017 challenge dataset involving 210 HGG and 75 LGG samples. The 2-fold cross-validation test method and seven metrics (accuracy, AUC, sensitivity, specificity, g-mean, precision, f-measure) are processed to evaluate the performance of frameworks efficiently. In experiments, the most effective framework is provided that uses FOS, data including three phase combinations, the Wilcoxon feature-ranking approach, and the GM-CPSO-NN method. Consequently, our framework achieved remarkable scores of 90.18% (accuracy), 85.62% (AUC), 95.24% (sensitivity), 76% (specificity), 85.08% (g-mean), 91.74% (precision), and 93.46% (f-measure) for HGG/LGG discrimination of 3D brain MRI data. Graphical abstract.

LDA Optimized Multi-scale Texture Features Based Diagnosis Method of Defects inside Insulated Tubular Busbars

In this paper, a method for diagnosing defects inside insulated tubular busbars based on LDA optimized multi-scale texture features is proposed to help to guarantee stable operation of the tubular busbars and the whole power grid. Firstly, multi-scale PRPD spectrum space was built with the UHF discharge signals of different defects by image pyramid theory. Then first-order, second-order and higher-order texture statistics were extracted from each image in the multi-scale PRPD spectrum space to form multi-scale texture features and LDA algorithm was used to optimize the features. The method was used to make texture features contain more information about partial discharge and help to improve the accuracy of diagnosis. Experiments were conducted on a 40.5kV insulated tubular busbar and CART classification trees were established as a classifier to identify the types of defects. The results of experiments show that this method can identify the defects of insulated tubular busbars accurately.

Investigation of poly(α-methyl styrene) tacticity synthesized by photo-polymerization

In this work, poly(α-methyl styrene) (PAS) was synthesized by free radical photo-polymerization. The tacticity and microstructure of PAS were studied by 13C nuclear magnetic resonance spectroscopy (13C NMR). Assignments of all stereo sequences were carried out at pentad and hexad levels of quaternary aliphatic, quaternary aromatic, and methylene carbons using 13C NMR in deuterated chloroform at 20 °C and 50 °C, respectively. It was found that by increasing the NMR acquisition temperature, higher resolution and higher splitting of the peaks were achieved for all mentioned carbons. Bernoullian and first-order Markov statistics were used for all carbons and the results were compared with experimental data by statistical method. It was shown that Bernoullian statics model fit slightly better than first-order Markov statics model for the assigned sequences. The results indicated that corresponding values of probability (Pm) was equal to 0.234, suggesting that the racemic addition was almost higher than the meso-one. The effect of the deuterated solvent on the peak resolution and the splitting of syndiotactic and isotactic sequences was investigated by replacing deuterated tetrahydrofuran with deuterated chloroform. It was observed that the peak resolution and the splitting in deuterated chloroform are much more than deuterated tetrahydrofuran.

Generalized entropies, density of states, and non-extensivity

The concept of entropy connects the number of possible configurations with the number of variables in large stochastic systems. Independent or weakly interacting variables render the number of configurations scale exponentially with the number of variables, making the Boltzmann–Gibbs–Shannon entropy extensive. In systems with strongly interacting variables, or with variables driven by history-dependent dynamics, this is no longer true. Here we show that contrary to the generally held belief, not only strong correlations or history-dependence, but skewed-enough distribution of visiting probabilities, that is, first-order statistics, also play a role in determining the relation between configuration space size and system size, or, equivalently, the extensive form of generalized entropy. We present a macroscopic formalism describing this interplay between first-order statistics, higher-order statistics, and configuration space growth. We demonstrate that knowing any two strongly restricts the possibilities of the third. We believe that this unified macroscopic picture of emergent degrees of freedom constraining mechanisms provides a step towards finding order in the zoo of strongly interacting complex systems.

Fault Diagnosis Based on Multiscale Texture Features of Cable Terminal on EMU of High-Speed Railway

The problem to identify the type of source generating discharges detected during campaigns of partial discharge (PD) measurements in cable terminal is of utmost importance in the quality control and state detection of the power supply system of a high-speed railway train. The purpose of this article is to provide a novel PD feature and a potential tool for cable terminal PD identification. The traditional PD type of high-speed train cable terminal is usually identified by PD test parameters, such as amplitude, PD times, and discharge energy. These features contain limited information and are easily affected by external factors, resulting in large numerical fluctuations, making the reliability of the test results low, which makes the operation, maintenance, and repair decision of high-speed train cables very difficult. Therefore, this article proposes a method that can extract rich feature information from PD data and distinguish PD types accurately and reliably by combining with a classification algorithm. First, to make full use of the characteristics of PD data, the single-scale PRPD images are transformed into multiscale PRPD images by using image pyramid theory. Then, the first-order texture statistics, the second-order texture statistics, and the higher order texture statistics are extracted from the multiscale PRPD space. Finally, the input features are optimized and classified by random forest and compared with other feature extraction algorithms. The results show that this method can identify PD type accurately and reliably, and the recognition accuracy is better than that of a single PD discharge feature.

Video person re-identification with global statistic pooling and self-attention distillation

Most existing methods for video person re-identification apply spatial-temporal global average or attention pooling to aggregate frame-level feature and get video-level feature. The obtained video-level feature models only the first-order statistics of the appearance feature from holistic video, resulting in limited representation capability of the feature network. In this paper, we propose a novel Global Statistic Pooling network (GSPnet) which takes full advantage of the second-order information for enhancing modeling capability. Firstly, a novel global statistic pooling module is proposed to summarize both the first- and second-order statistics across frame-level feature, and then transfer them into a compact and robust video-level feature embedding. Secondly, a statistic-based attention block is incorporated into multiple stages of convolutional networks to fully explore the second-order representations from low- to high-level features. To enhance the representation learning ability and further boost re-identification (re-ID) performance, we also propose a multi-level self-attention distillation training scheme, which squeezes the knowledge learned in the deeper portion of the networks into the shallow ones. Extensive experimental results have demonstrated the effectiveness and superiority of our approach on four popular video person re-ID datasets.

A Low-Complexity Beamforming Design for Multiuser Wireless Energy Transfer

Wireless energy transfer (WET) is a green enabler of low-power Internet of Things (IoT). Therein, traditional optimization schemes relying on full channel state information (CSI) are often too costly to implement due to excessive energy consumption and high processing complexity. This letter proposes a simple, yet effective, energy beamforming scheme that allows a multi-antenna power beacon (PB) to fairly power a set of IoT devices by only relying on the first-order statistics of the channels. In addition to low complexity, the proposed scheme performs favorably as compared to benchmarking schemes and its performance improves as the number of PB’s antennas increases. Finally, it is shown that further performance improvement can be achieved through proper angular rotations of the PB.

Application of the Mixture of Lognormal Distribution to Represent the First-Order Statistics of Wireless Channels

In a wireless channel, shadow fading along with multipath effect causes random fluctuations of the received signal strength at the receiver. Several composite fading distributions are available for modeling the randomness. Among them, Lognormal-based models are unable to give a closed-form expression of the composite distribution, but Gamma-based model overcomes the problem faced by the former one. However, the Gamma-based model comes with a complicated mathematical function. To get rid of that, mixture of Lognormal (MoLN) distribution is used in this article. Maximum-likelihood approach is used for parameters estimation with the help of the expectation maximization algorithm. Simulation results have been provided in order to show the accuracy of the proposed mixture model. The first-order statistics of both the envelope and signal-to-noise ratio of a wireless channel are represented by a simple mathematical form. Several performance metrics, used in a wireless system, are expressed in a closed form for the MoLN distribution. The fitness of MoLN distribution is tested. Quantitatively, by measuring metrics like mean square error and Kullback–Leibler divergence, it has been shown that MoLN distribution is a suitable candidate to represent the composite wireless channel accurately.

Multi-scale structural kernel representation for object detection

Existing high-performance object detection methods greatly benefit from the powerful representation ability of deep convolutional neural networks (CNNs). Recent researches show that integration of high-order statistics remarkably improves the representation ability of deep CNNs. However, high-order statistics for object detection lie in two challenges. Firstly, previous methods insert high-order statistics into deep CNNs as global representations, which lose spatial information of inputs, and so are not applicable to object detection. Furthermore, high-order statistics have special structures, which should be considered for proper use of high-order statistics. To overcome above challenges, this paper proposes a Multi-scale Structural Kernel Representation (MSKR) for improving performance of object detection. Our MSKR is developed based on the polynomial kernel approximation, which does not only draw into high-order statistics but also preserve the spatial information of input. To consider geometry structures of high-order representations, a feature power normalization method is introduced before computation of kernel representation. Comparing with the most commonly used first-order statistics in existing CNN-based detectors, our MSKR can generate more discriminative representations, and so be flexibly integrated into deep CNNs for improving performance of object detection. By adopting the proposed MSKR to existing object detection methods (i.e., Faster R-CNN, FPN, Mask R-CNN and RetinaNet), it achieves clear improvement on three widely used benchmarks, while obtaining very competitive performance with state-of-the-art methods.

Classification of glaucoma using hybrid features with machine learning approaches

Glaucoma also known as silent theft of sight, is the prevailing source of sightedness across the globe. It is mainly caused owing to damage in the optical nerve of an eye leading to permanent blindness. Traditional approaches used by the ophthalmologists for diagnosis include assessment of intraocular pressure by tonometry, pachymetry etc. But all these evaluations are time-consuming, require human interaction and may be prone to subjective errors. Thus, to overcome these challenges, researchers are working in the field of medical imaging by analysis of the retinal images for glaucoma diagnosis. Also, Computer aided diagnosis (CAD) systems can be developed to overcome these challenges using machine learning approaches for classifying retinal images as ‘abnormal’ and ‘normal’. This paper presents a new set of reduced hybrid features derived from structural and nonstructural features to classify the retinal fundus image, which could serve as a second opinion for ophthalmologists. The structural features extracted include Disc damage likelihood scale (DDLS) and Cup to disc ratio (CDR). Whereas, non-structural features include Grey level run length matrix (GLRM), Grey level co-occurrence matrix (GLCM), First order statistical (FoS), Higher order spectra (HOS), Higher order cumulant (HOC) and Wavelets. Finally, the paper presents a comparative analysis of K nearest neighbor (k-NN), Neural network (NN), Random forest (RF), Support vector machine (SVM) and Naïve bayes (NB) using metrics such as accuracy, specificity, precision and sensitivity.

Enhancing the differentiation of pulmonary lymphoma and fungal pneumonia in hematological patients using texture analysis in 3-T MRI

ObjectivesTo evaluate texture analysis in nonenhanced 3-T MRI for differentiating pulmonary fungal infiltrates and lymphoma manifestations in hematological patients and to compare the diagnostic performance with that of signal intensity quotients (“nonenhanced imaging characterization quotients,” NICQs).MethodsMR scans were performed using a speed-optimized imaging protocol without an intravenous contrast medium including axial T2-weighted (T2w) single-shot fast spin-echo and T1-weighted (T1w) gradient-echo sequences. ROIs were drawn within the lesions to extract first-order statistics from original images using HeterogeneityCAD and PyRadiomics. NICQs were calculated using signal intensities of the lesions, muscle, and fat. The standard of reference was histology or clinical diagnosis in follow-up. Statistical testing included ROC analysis, clustered ROC analysis, and DeLong test. Intra- and interrater reliability was tested using intraclass correlation coefficients (ICC).ResultsThirty-three fungal infiltrates in 16 patients and 38 pulmonary lymphoma manifestations in 19 patients were included. Considering the leading lesion in each patient, diagnostic performance was excellent for T1w entropy (AUC 80.2%; p < 0.005) and slightly inferior for T2w energy (79.9%; p < 0.005), T1w uniformity (79.6%; p < 0.005), and T1w energy (77.0%; p < 0.01); the best AUC for NICQs was 72.0% for T2NICQmean (p < 0.05). Intra- and interrater reliability was good to excellent (ICC > 0.81) for these parameters except for moderate intrarater reliability of T1w energy (ICC = 0.64).ConclusionsT1w entropy, uniformity, and energy and T2w energy showed the best performances for differentiating pulmonary lymphoma and fungal pneumonia and outperformed NICQs. Results of the texture analysis should be checked for their intrinsic consistency to identify possible incongruities of single parameters.Key Points• Texture analysis in nonenhanced pulmonary MRI improves the differentiation of pulmonary lymphoma and fungal pneumonia compared with signal intensity quotients.• T1w entropy, uniformity, and energy along with T2w energy show the best performances for differentiating pulmonary lymphoma from fungal pneumonia.• The results of the texture analysis should be checked for their intrinsic consistency to identify possible incongruities of single parameters.

Frequency Scaling Model for the Prediction of Total Tropospheric Attenuation Time Series at EHF

A frequency scaling model aiming to predict total tropospheric attenuation time series at EHF is presented. One version of the model [total attenuation frequency scaling (TAFS)] is more accurate but, besides the reference attenuation time series at low frequency, it also requires additional ancillary inputs (e.g., radiometric data and information on the raindrop size distribution); a second version (S-TAFS) is simpler, though at the expenses of a slightly worse performance. Both TAFS and simplified TAFS (S-TAFS) first rely on isolating the attenuation induced by different atmospheric constituents, each of which is separately up-scaled to the target frequency. The two methods are tested against a full year of data collected at Ka -band and $Q$ -band at Politecnico di Milano, Milan, Italy, in the framework of the Alphasat Aldo Paraboni propagation experiment. Results indicate that both methods offer a very good accuracy in scaling the total attenuation from Ka- to $Q$ -bands, both in terms of first-order statistics and of time series. This corroborates the use of TAFS and S-TAFS to predict the total tropospheric attenuation at much higher frequency bands (e.g., the $W$ -band), for which no measurements are currently available, starting from the largely available Ka -band measurements.