High Frequency Components Research Articles

Improved filtering methods to suppress cardiovascular contamination in electrical impedance tomography recordings

Objective. Electrical impedance tomography (EIT) produces clinical useful visualization of the distribution of ventilation inside the lungs. The accuracy of EIT-derived parameters can be compromised by the cardiovascular signal. Removal of these artefacts is challenging due to spectral overlapping of the ventilatory and cardiovascular signal components and their time-varying frequencies. We designed and evaluated advanced filtering techniques and hypothesized that these would outperform traditional low-pass filters. Approach. Three filter techniques were developed and compared against traditional low-pass filtering: multiple digital notch filtering (MDN), empirical mode decomposition (EMD) and the maximal overlap discrete wavelet transform (MODWT). The performance of the filtering techniques was evaluated (1) in the time domain (2) in the frequency domain (3) by visual inspection. We evaluated the performance using simulated contaminated EIT data and data from 15 adult and neonatal intensive care unit patients. Main result. Each filter technique exhibited varying degrees of effectiveness and limitations. Quality measures in the time domain showed the best performance for MDN filtering. The signal to noise ratio was best for DLP, but at the cost of a high relative and removal error. MDN outbalanced the performance resulting in a good SNR with a low relative and removal error. MDN, EMD and MODWT performed similar in the frequency domain and were successful in removing the high frequency components of the data. Significance. Advanced filtering techniques have benefits compared to traditional filters but are not always better. MDN filtering outperformed EMD and MODWT regarding quality measures in the time domain. This study emphasizes the need for careful consideration when choosing a filtering approach, depending on the dataset and the clinical/research question.

Analytical solution to a coupled system including tuned liquid damper and single degree of freedom under free vibration with modal decomposition method

This research focuses on employing a linear analytical approach to transform free surface waves and velocities into mode coordinates, with the aim of investigating the free vibration behavior of a coupled system consisting of a Single Degree of Freedom and a sloshing tank. Through a series of manipulations and simplifications of the coupled equations, a fourth-order ordinary differential equation is derived to showcase the overall response of the system, highlighting the contribution of each odd mode. Key concepts explored include system stability, mode-specific natural periods, establishment of initial boundary conditions, and formulation of the complete system response. The analytical method applied to study Tuned Liquid Dampers, a type of elevated sloshing tank, reveals that in higher modes, the lower frequency aligns with the structural natural frequency, while the higher frequency is approximately n times the structural natural frequency (where n is the odd mode number). This approach also elucidates why the system's response does not exhibit a higher-frequency component in higher modes. The study further investigates concepts such as employing an initial perturbation to excite higher frequencies and the potential for approximating the system through the first mode. Additionally, a numerical model was developed using variable separation and modal decomposition methods to complement and validate the analytical approach. Finally, further verification of the model was performed using the Preismann scheme applied to the relevant equations and the central upwind applied to nonlinear equations.

A weight induced contrast map for infrared and visible image fusion

Fusion involves merging details from both infrared (IR) and visible images to generate a unified composite image that offers richer and more valuable information than either individual image. Surveillance, navigation, remote sensing, and military applications require various imaging modalities, including visible and IR, to oversee specific scenes. These sensors provide supplementary data and improve situational understanding, so it is essential to fuse this information into a single image. Fusing IR and visible images presents several challenges due to the differences in imaging modalities, data characteristics, and the need for accurate and meaningful integration of information. In this context, a novel image fusion architecture focuses on enhancing prominent targets, with the objective of integrating thermal information from infrared images into visible images while preserving textural details within the visible images. Initially, in the proposed algorithm, the images from different sensors are divided into components of high and low frequencies using a Guided filter and an Average filter, respectively. A unique contrast detection mechanism is proposed that is capable of preserving the contrast information from the original images. Further, the contrast details of the IR and visible images are enhanced using local standard deviation filtering and local range filtering, respectively. We have developed a new weight map construction strategy that can effectively preserve the supplemental data from both the original images. These weights and gradient details of the source images are utilized to preserve the salient feature details of the images acquired from the various modalities. A decision-making approach is utilized among the high-frequency components of the original images to retain the prominent feature details of the source images. Finally, the salient feature details and the prominent feature details are integrated to generate the fused image. The developed technique is validated using both subjective and quantitative perspectives. The developed approaches provide EN, MI, Nabf, and SD of 6.86815, 13.73269, 0.15390, and 78.16158 respectively against deep learning-based approaches. Also, the proposed algorithm provides EN, MI, Nabf, FMIw, and Qabf against 6.86815, 13.73269, 0.15390, 0.41634 and 0.47196 respectively against existing traditional fusion methods. It is observed that the developed technique provides adequate accuracy against twenty-seven state-of-the-art techniques.

A comprehensive multivariate analysis of the center of pressure during quiet standing in patients with Parkinson's disease

BackgroundWe hypothesized that postural instability observed in individuals with Parkinson's disease (PD) can be classified as distinct subtypes based on comprehensive analyses of various evaluated parameters obtained from time-series of center of pressure (CoP) data during quiet standing. The aim of this study was to characterize the postural control patterns in PD patients by performing an exploratory factor analysis and subsequent cluster analysis using CoP time-series data during quiet standing.Methods127 PD patients, 47 aged 65 years or older healthy older adults, and 71 healthy young adults participated in this study. Subjects maintain quiet standing for 30 s on a force platform and 23 variables were calculated from the measured CoP time-series data. Exploratory factor analysis and cluster analysis with a Gaussian mixture model using factors were performed on each variable to classify subgroups based on differences in characteristics of postural instability in PD.ResultsThe factor analysis identified five factors (magnitude of sway, medio-lateral frequency, anterio-posterior frequency, component of high frequency, and closed-loop control). Based on the five extracted factors, six distinct subtypes were identified, which can be considered as subtypes of distinct manifestations of postural disorders in PD patients. Factor loading scores for the clinical classifications (younger, older, and PD severity) overlapped, but the cluster classification scores were clearly separated.ConclusionsThe cluster categorization clearly identifies symptom-dependent differences in the characteristics of the CoP, suggesting that the detected clusters can be regarded as subtypes of distinct manifestations of postural disorders in patients with PD.

Finite amplitude wave propagation through bubbly fluids

The existence of only a few bubbles could drastically reduce the acoustic wave speed in a liquid. Wood’s equation models the linear sound speed, while the speed of an ideal shock waves is derived as a function of the pressure ratio across the shock. The common finite amplitude waves lie, however, in between these limits. We show that in a bubbly medium, the high frequency components of finite amplitude waves are attenuated and dissipate quickly, but a low frequency part remains. This wave is then transmitted by the collapse of the bubbles and its speed decreases with increasing void fraction. We demonstrate that the linear and the shock wave regimes can be smoothly connected through a Mach number based on the collapse velocity of the bubbles.

Enhanced wavelet based spatiotemporal fusion networks using cross-paired remote sensing images

Spatiotemporal fusion can provide remote sensing images with both high temporal and high spatial resolution for earth observation applications. Most of the state-of-the-art models require three or even five images as input, which may lead to difficulties in practical applications due to bad weather or data missing. In this paper, the enhanced cross-paired wavelet based spatiotemporal fusion networks (ECPW-STFN) are proposed to improve the accuracy and quality of the fusions with fewer remote sensing images as inputs. Wavelet transform is introduced into spatiotemporal image fusion to achieve separate training of the high and low frequency components of the image to better extract features of different levels. In addition, a compound loss function containing wavelet loss is proposed to facilitate the preservation of details. Also, an enhancement module with convolutional block attention is put forward to further refine the prediction quality. Experiments were conducted to compare the proposed ECPW-STFN with five state-of-the-art methods on the public CIA and Daxing datasets. The results show ECPW-STFN is better than GAN-STFM which also requires two images, and not inferior to the methods requiring at least three inputs, even exceeds the optimal MLFF-GAN in some cases, proving its great superiority and potential. The code will be available at https://github.com/lixinghua5540/ECPW-STFN.

Millijoule Terahertz Radiation from Laser Wakefields in Nonuniform Plasmas.

We report the experimental measurement of millijoule terahertz (THz) radiation emitted in the backward direction from laser wakefields driven by a femtosecond laser pulse of few joules interacting with a gas target. By utilizing frequency-resolved energy measurement, it is found that the THz spectrum exhibits two peaks located at about 4.5 and 9.0THz, respectively. In particular, the high frequency component emerges when the drive laser energy exceeds 1.26J, at which electron acceleration in the forward direction is detected simultaneously. Theoretical analysis and particle-in-cell simulations indicate that the THz radiation is generated via mode conversion from the laser wakefields excited in plasma with an up-ramp profile, where radiations both at the local electron plasma frequency and its harmonics are produced. Such intense THz sources may find many applications in ultrafast science, e.g., manipulating the transient states of matter.

A novel chaotic image encryption is based on fractional wavelet decomposition and quantum transform model

A new chaotic image encryption scheme based on fractional order wavelet decomposition is proposed in this paper. Initially, the image undergoes a three-stage fractional-order wavelet decomposition, resulting in high-frequency and low-frequency components. High-frequency components are dislocated using a chaotic system, while low-frequency components are dislocated employing a fractional matrix model. Subsequently, the scrambled image undergoes quantum encoding followed by the Arnold transform to yield the final scrambled result. Ultimately, the encryption process involves iteratively acquiring the M4 key matrix and diffusing the quantum image through permutation to obtain the final encryption result. Simulation experiments and numerical analyses demonstrate the high security level of the proposed encryption method.

Bolt Looseness Detection in Brake Disc of High-Speed Rail Using Wavelet Packet Decomposition and 1D Convolutional Neural Network

The effective and in-time detection of bolt looseness in the brake disc of high-speed rail is of great significance to ensure the safe operation of the train. The hammer tapping method is one of the most widely used methods for bolt looseness detection. Due to the complex structure of the brake disc, the vibration characteristics associated with bolt looseness are coupled with the transmission path from the head of the bolt through the brake disc to the sensors, which make it difficult to quantitively identify the bolt looseness. In this work, a method based on wavelet packet decomposition and a one-dimensional convolutional neural network (1D CNN) is proposed to quantitively detect the bolt looseness of brake disc. Firstly, the vibration signals collected from the two sensors mounted on the nut side are fused by autocorrelation summation to reduce the influence of transmission path. Secondly, the fused signal is decomposed to sub signals in low frequency and high frequency components by wavelet packet decomposition. The relative difference of wavelet packet energy among sub signals is extracted as the features to enhance the difference among different degrees of looseness. Finally, the 1D CNN model is established and trained by the features of energy relative difference to quantitively identify the bolt looseness. To validate the effectiveness of the proposed method, an experimental platform for bolt looseness detection in brake disc is constructed. Compared with the single-channel 1D CNN and fused signal-1D CNN models, the accuracy of the proposed method is approximately 97%, which confirms the superiority of the proposed method.

Learning Generalized Segmentation for Foggy-Scenes by Bi-directional Wavelet Guidance

Learning scene semantics that can be well generalized to foggy conditions is important for safety-crucial applications such as autonomous driving. Existing methods need both annotated clear images and foggy images to train a curriculum domain adaptation model. Unfortunately, these methods can only generalize to the target foggy domain that has seen in the training stage, but the foggy domains vary a lot in both urban-scene styles and fog styles. In this paper, we propose to learn scene segmentation well generalized to foggy-scenes under the domain generalization setting, which does not involve any foggy images in the training stage and can generalize to any arbitrary unseen foggy scenes. We argue that an ideal segmentation model that can be well generalized to foggy-scenes need to simultaneously enhance the content, de-correlate the urban-scene style and de-correlate the fog style. As the content (e.g., scene semantic) rests more in low-frequency features while the style of urban-scene and fog rests more in high-frequency features, we propose a novel bi-directional wavelet guidance (BWG) mechanism to realize the above three objectives in a divide-and-conquer manner. With the aid of Haar wavelet transformation, the low frequency component is concentrated on the content enhancement self-attention, while the high frequency component is shifted to the style and fog self-attention for de-correlation purpose. It is integrated into existing mask-level Transformer segmentation pipelines in a learnable fashion. Large-scale experiments are conducted on four foggy-scene segmentation datasets under a variety of interesting settings. The proposed method significantly outperforms existing directly-supervised, curriculum domain adaptation and domain generalization segmentation methods. Source code is available at https://github.com/BiQiWHU/BWG.

EEG stress classification based on Doppler spectral features for ensemble 1D-CNN with LCL activation function

The paper proposes an induced stress classification algorithm that uses features from the Doppler spectrum. In this approach, a reference signal source is used to obtain the quadrature and in-phase components of the EEG signal. The higher frequency components from the in-phase and quadrature are then eliminated using a pair of low-pass filters. The Doppler spectrum was then constructed from which the Doppler frequency features are then estimated. The features that were thus obtained are trained using an ensemble 1D-CNN (one-dimensional Convolutional neural network) which uses two sections of 1D-CNN. The first section 1D-CNN trains the features based on the EEG signal classes namely Stroop test, arithmetic task, and mirror tasks, while the second 1D-CNN section trains the features based on the EEG signal intense classes namely high, low, and medium stress. We also propose a linear-cosine-linear (LCL) activation function for the ensemble 1D-CNN which was derived from the cosine signal. The proposed stress classification scheme was evaluated using the SAM-40 datasets with induced stress classes namely arithmetic task, Stroop color-word test, and mirror image recognition task with stress levels namely high, low, and medium with the evaluation metrics such as precision, F1-score, accuracy, specificity, and recall. The proposed stress classification scheme attains an average accuracy, precision, and recall of 95.25%, 95.22%, and 92.9% when evaluated in 9 classes of EEGs.

Research on interference compensation methods for color image sensors based on iterative learning

PurposeFixed mode noise and random mode noise always exist in the image sensor, which affects the imaging quality of the image sensor. The charge diffusion and color mixing between pixels in the photoelectric conversion process belong to fixed mode noise. This study aims to improve the image sensor imaging quality by processing the fixed mode noise.Design/methodology/approachThrough an iterative training of an ergoable long- and short-term memory recurrent neural network model, the authors obtain a neural network model able to compensate for image noise crosstalk. To overcome the lack of differences in the same color pixels on each template of the image sensor under flat-field light, the data before and after compensation were used as a new data set to further train the neural network iteratively.FindingsThe comparison of the images compensated by the two sets of neural network models shows that the gray value distribution is more concentrated and uniform. The middle and high frequency components in the spatial spectrum are all increased, indicating that the compensated image edges change faster and are more detailed (Hinton and Salakhutdinov, 2006; LeCun et al., 1998; Mohanty et al., 2016; Zang et al., 2023).Originality/valueIn this paper, the authors use the iterative learning color image pixel crosstalk compensation method to effectively alleviate the incomplete color mixing problem caused by the insufficient filter rate and the electric crosstalk problem caused by the lateral diffusion of the optical charge caused by the adjacent pixel potential trap.

Capacitance Analysis for Volume Reduction Based on Integrated Buck and Buck-Boost LED Driver

This paper presents a design methodology to improve the power density of integrated converters operating at universal input voltage fulfilling the standards of connection to the grid and load. The proposed integrated two-stage converter is composed of the power factor correction (PFC) stage behaving as a current source to the bus, and the power control (PC) stage which provides continuous energy to the LED load. Traditionally, the low-frequency ripple (LFR) filtering process is performed by the bus capacitor placed in the output of the PFC stage, while the PC stage output capacitor works only as a high-frequency (HF) filter. The idea is to explore the lower and fixed operating voltage characteristics of the PC stage to share the LFR filtering with the PFC stage. Thus, a mathematical analysis is carried out, considering the influence of the LED characteristics, bus voltage, and capacitances to predict the LFR in the LED current. A case study, composed of an integrated buck and buck-boost converter to supply a 75W LED load, is presented. For the traditional design method, the LED driver needs an 820uF/160V PFC bus capacitor and a 10uF/80V output capacitor to filter the HF components, representing a total capacitors' volume of 17.9cm³. With the proposed analysis, the optimized driver circuit requires a 220uF/160V PFC bus capacitor and a 470uF/80V output capacitor, resulting in a total capacitors' volume of 9.7cm³, providing a volume reduction of approximately 45%.

Sensorimotor content of multi-unit activity recorded in the paramedian lobule of the cerebellum using carbon fiber microelectrode arrays.

The cerebellum takes in a great deal of sensory information from the periphery and descending signals from the cerebral cortices. It has been debated whether the paramedian lobule (PML) in the rat and its paravermal regions that project to the interpositus nucleus (IPN) are primarily involved in motor execution or motor planning. Studies that have relied on single spike recordings in behaving animals have led to conflicting conclusions regarding this issue. In this study, we tried a different approach and investigated the correlation of field potentials and multi-unit signals recorded with multi-electrode arrays from the PML cortex along with the forelimb electromyography (EMG) signals in rats during behavior. Linear regression was performed to predict the EMG signal envelopes using the PML activity for various time shifts (±25, ±50, ±100, and ± 400 ms) between the two signals to determine a causal relation. The highest correlations (~0.5 on average) between the neural and EMG envelopes were observed for zero and small (±25 ms) time shifts and decreased with larger time shifts in both directions, suggesting that paravermal PML is involved both in processing of sensory signals and motor execution in the context of forelimb reaching behavior. EMG envelopes were predicted with higher success rates when neural signals from multiple phases of the behavior were utilized for regression. The forelimb extension phase was the most difficult to predict while the releasing of the bar phase prediction was the most successful. The high frequency (>300 Hz) components of the neural signal, reflecting multi-unit activity, had a higher contribution to the EMG prediction than did the lower frequency components, corresponding to local field potentials. The results of this study suggest that the paravermal PML in the rat cerebellum is primarily involved in the execution of forelimb movements rather than the planning aspect and that the PML is more active at the initiation and termination of the behavior, rather than the progression.

Acute combined effects of concurrent physical activities on autonomic nervous activation during cognitive tasks.

Backgrounds: The validity of heart rate variability (HRV) has been substantiated in mental workload assessments. However, cognitive tasks often coincide with physical exertion in practical mental work, but their synergic effects on HRV remains insufficiently established. The study aims were to investigate the combined effects of cognitive and physical load on autonomic nerve functions. Methods: Thirty-five healthy male subjects (aged 23.5 ± 3.3years) were eligible and enrolled in the study. The subjects engaged in n-back cognitive tasks (1-back, 2-back, and 3-back) under three distinct physical conditions, involving isotonic contraction of the left upper limb with loads of 0kg, 3kg, and 5kg. Electrocardiogram signals and cognitive task performance were recorded throughout the tasks, and post-task assessment of subjective experiences were conducted using the NASA-TLX scale. Results: The execution of n-back tasks resulted in enhanced perceptions of task-load feelings and increased reaction times among subjects, accompanied by a decline in the accuracy rate (p < 0.05). These effects were synchronously intensified by the imposition of physical load. Comparative analysis with a no-physical-load scenario revealed significant alterations in the HRV of the subjects during the cognitive task under moderate and high physical conditions. The main features were a decreased power of the high frequency component (p < 0.05) and an increased low frequency component (p < 0.05), signifying an elevation in sympathetic activity. This physiological response manifested similarly at both moderate and high physical levels. In addition, a discernible linear correlation was observed between HRV and task-load feelings, as well as task performance under the influence of physical load (p < 0.05). Conclusion: HRV can serve as a viable indicator for assessing mental workload in the context of physical activities, making it suitable for real-world mental work scenarios.

Time-dependent hormesis transfer from five high-frequency personal care product components to mixtures

There is potential for personal care products (PCPs) components and mixtures to induce hormesis. How hormesis is related to time and transmitted from components to mixtures are not clear. In this paper, we conducted determination of components in 16 PCP products and then ran frequent itemset mining on the component data. Five high-frequency components (HFCs), betaine (BET), 1,3-butanediol (BUT), ethylenediaminetetraacetic acid disodium salt (EDTA), glycerol (GLO), and phenoxyethanol (POE), and 14 mixtures were identified. For each mixture system, one mixture ray with the actual mixture ratios in the products was selected. Time-dependent microplate toxicity analysis was used to test the luminescence inhibition toxicity of five HFCs and 14 mixture rays to Vibrio qinghaiensis sp.-Q67 at 12 concentration gradients and eight exposure times. It is showed that BET, EDTA, POE, and 13 mixture rays containing at least one J-type component showed time-dependent hormesis. Characteristic parameters used to describe hormesis revealed that the absolute value of the maximum stimulatory effect (|Emin|) generally increased with time. Notably, mixtures composed of POE and S-type components showed greater |Emin| than POE alone at the same time. Importantly, the maximum stimulatory effective concentration, NOEC/the zero effective concentration point, and EC50 remained relatively stable. Nine hormesis transmission phenomena were observed in different mixture rays. While all mixtures primarily exhibited additive action, varying degrees of synergism and antagonism were noted in binary mixtures, with no strong synergism or antagonism observed in ternary and quaternary mixtures. These findings offer valuable insights for the screening of HFCs and their mixtures, as well as the study of hormesis transmission in personal care products.

High-time-resolution analysis of meridional tides in the upper mesosphere and lower thermosphere at mid-latitudes measured by the Falkland Islands SuperDARN radar

Abstract. Solar tides play a major role in the dynamics of the upper mesosphere and lower thermosphere (MLT). Hence, a comprehensive understanding of these tides is important for successful modelling of the MLT region. Most ground-based observations of tidal variations in the MLT have been from meteor radar measurements with a temporal resolution of 1 h. Here, we take a different perspective on these tidal variations using high-resolution 1 min neutral-wind measurements from the Falkland Islands SuperDARN (Super Dual Auroral Radar Network) radar. This analysis shows that these higher-resolution data can be used to identify higher frequency tidal components than are typically observed by meteor radars (up to a heptadiurnal component). It also shows evidence of significant power in these higher frequency components, particularly in the quaddiurnal component, which may be particularly suitable for a global analysis using high-resolution SuperDARN neutral-wind measurements. The high-resolution analysis also shows evidence of fluctuations with a frequency of 1.5 cycles per day, as well as higher-frequency fluctuations, accompanying a quasi-2 d (two day) wave. We discuss the limitations of this high-resolution analysis method and the new opportunities that it may provide. We conclude that higher-resolution SuperDARN neutral-wind measurements need to be better exploited in the future as they provide a complementary way of studying tides and waves in the MLT.

Hybrid simulation of instabilities in capacitively coupled RF CF4/Ar plasmas driven by a dual frequency source

Instabilities in capacitively coupled Ar/CF4 plasma discharges driven by dual frequency sources are investigated using a one-dimensional fluid/electron Monte Carlo hybrid model. Periodic oscillations of the electron density and temperature on the timescale of multiple low frequency (LF) periods are observed. As the electron density increases, an intense oscillation of the electron temperature within each high frequency (HF) period is initiated. This causes a fluctuation of the electron density and results in a discharge instability. This phenomenon is consistent with the discharge behavior observed in scenarios with single-frequency (SF) sources, as reported by Dong et al (2022 Plasma Sources Sci. Technol. 31 025006). However, unlike the SF case, plasma parameters such as the electron density, electric field, electron power absorption and ionization rate exhibit not only periodic fluctuations but also a spatial asymmetry under the influence of the dual-frequency source. This spatial asymmetry leads to a non-uniform distribution of the electron density between the electrodes, which is related to a spatially asymmetric electric field, electron heating, and ionization around a region of minimum electron density (inside the bulk). This region of minimum electron density is shifted back and forth through the entire plasma bulk from one electrode to the other within multiple LF period. The above phenomena are related to superposition effect between the instabilities and the dual-frequency source. Moreover, the time averaged electric field influences the spatio-temporal evolution of ion fluxes. The ion fluxes at the electrodes, which play an important role in etching processes, are affected by both the high and LF components of the driving voltage waveform as well as the observed instabilities. As the HF increases, the electronegativity and electron temperature are reduced and the electron density increases, resulting in a gradual disappearance of the instabilities.

Impact of COVID-19-associated Dysautonomia Phenotypes on the Effectiveness of Venovenous Extracorporeal Membrane Oxygenation

RELEVANCE. It is known that patients with severe cases of the novel coronavirus infection (COVID-19) are characterized by the development of COVID-19-associated dysautonomia (COVID-19-DA). At the moment, there are no studies examining the impact of this phenomenon on the course and outcomes of the disease in the most severe cohort of patients with COVID-19, namely those requiring venovenous extracorporeal membrane oxygenation (VV ECMO). AIM OF STUDY. The purpose of the presented work is to study the effect of different COVID-19-DA phenotypes on the performance parameters and effectiveness of VV ECMO, gas exchange and hemodynamics in patients with COVID-19. MATERIAL AND METHODS. The study included 20 patients, 12 (60%) women, 8 (40%) men, with COVID-19 who underwent VV ECMO. Patients underwent 24-hour Holter monitoring with assessment of the spectral parameters of heart rate variability (HRV): the low-frequency component (LF), the high-frequency component (HF), the ratio of the low-frequency component to the high frequency component (LF / HF) on days 1, 3, 5 of the VV ECMO. Patients were divided into three groups based on the identified COVID-19-DA phenotypes. The groups were compared in terms of gas exchange, hemodynamics, and VV ECMO performance parameters. RESULTS. The level of partial pressure of carbon dioxide in arterial blood (paCO2) in the phenotype with low sympathetic tone and high tone of the parasympathetic division of the autonomic nervous system (ANS) (lShP) was significantly higher than in the phenotype with normal sympathetic tone and high tone of the parasympathetic division of the ANS (nShP) with equal VV ECMO performance parameters. The heart rate (HR) in the nShP phenotype was significantly lower than in the lShP phenotype. A significant increase in respiratory dysfunction was revealed over time in the lShP phenotype. Weaning from VV ECMO in the nShP phenotype was successful in 50 %, whereas in the lShP phenotype, weaning from VV ECMO was observed in 7,2 % of patients. No significant differences in the mortality rate were obtained. The most common cause of death in both groups was septic shock. CONCLUSIONS. The COVID-19-DA phenotype, manifested by decreased tone of the sympathetic division and increased tone of the parasympathetic division of the ANS, leads to low efficiency of VV ECMO, resulting in a statistically significantly less frequent ECMO discontinuation in those patients.

High Frequency Component Enhancement Network for Image Manipulation Detection

With the support of deep neural networks, the existing image manipulation detection (IMD) methods can detect manipulated regions within a suspicious image effectively. In general, manipulation operations (e.g., splicing, copy-move, and removal) tend to leave manipulation artifacts in the high-frequency domain of the image, which provides rich clues for locating manipulated regions. Inspired by this phenomenon, in this paper, we propose a High-Frequency Component Enhancement Network, short for HFCE-Net, for image manipulation detection, which aims to fully explore the manipulation artifacts left in the high-frequency domain to improve the localization performance in IMD tasks. Specifically, the HFCE-Net consists of two parallel branches, i.e., the main stream and high-frequency auxiliary branch (HFAB). The HFAB is introduced to fully explore high-frequency artifacts within manipulated images. To achieve this goal, the input image of the HFAB is filtered out of the low-frequency component by the Sobel filter. Furthermore, the HFEB is supervised with the edge information of the manipulated regions. The main stream branch takes the RGB image as input, and aggregates the features learned from the HFAB by the proposed multi-layer fusion (MLF) in a hierarchical manner. We conduct extensive experiments on widely used benchmarks, and the results demonstrate that our HFCE-Net exhibits a strong ability to capture high-frequency information within the manipulated image. Moreover, the proposed HFCE-Net achieves comparable performance (57.3%, 90.9%, and 73.8% F1 on CASIA, NIST, and Coverage datasets) and achieves 1.9%, 9.0%, and 1.5% improvement over the existing method.