Search

Background: Hepatic fibrosis is a common pathological process in the process of various liver injury and repair. It can develop into advanced cirrhosis or even liver cancer under the stimulation of various factors. Early diagnosis and intervention can significantly improve liver function and even reverse the histopathological process. Accurate non-invasive identification and staging of early liver fibrosis has become the research goal of many scholars. Objective: This study aimed to propose a model training method and system for liver fibrosis lesion detection based on CNN-Former. Study design: We first collect the clinical patient data and introduce the setup and data preprocessing of the experimental data, then introduce the experimental process of the feature engineering , the ablation experiment conducted and the evaluation index . Finally, compared this method with traditional method . Results: By modeling the physiological and biochemical characteristics of patients with liver fibrosis, the pre-detection of liver fibrosis is carried out with high efficiency, high quality and high standard, providing a solution for existing research to rely on image information. In addition, on the basis of the existing CNN-form model, multi-CNN-form adds a multi-task mechanism to enable the model to predict liver fibrosis and related complications at the same time, providing a solution to the problem that traditional models focus on specific diseases and poor transplantation. Also, the accuracy, recall, AUC value, F1 value and other evaluation indexes of the model were calculated based on the test set and confusion matrix to compare the effects of various models in predicting liver fibrosis pre-detection. Finally, the effectiveness of this method compared with the traditional method is demonstrated from many angles. This makes it more suitable in practical application scenarios. Conclusions: This paper mainly introduces a liver fibrosis pre-detection method based on deep learning, that is, the liver fibrosis pre-detection method based on CNN-former, which provides a solution for the existing research to rely on imaging information. In addition, the pre-detection method of liver fibrosis based on multi-CNN-former provides a solution to the problem of focusing on specific diseases and poor transplantation of traditional models.

Nickel-free high-manganese austenitic Fe–24.4Mn–4.04Al–0.057C steel was produced by smelting, and the homogenized forged billet was hot-rolled. The plastic deformation mechanism was investigated through tensile testing of the hot-rolled sample. Different characterization techniques such as scanning electron microscopy, transmission electron microscopy, electron backscattered diffraction, and X-ray diffraction were used to analyze the microstructural evolution of steel under different strain levels. The steel had a single austenite phase, which was stable during deformation. After hot rolling, annealing twins were observed in the microstructure of the steel. The steel showed an excellent combination of mechanical properties, like a tensile strength of 527 MPa, impact energy of 203 J at − 196 °C, and an elongation of 67% till fracture. At the initial deformation stage, the dislocations were generated within the austenite grains, entangled and accumulated at the grain boundaries and annealing twin boundaries. Annealing twins participated in plastic deformation and hindered the dislocation movement. As the deformation progressed, the dislocation slip was hindered and produced stress concentration, and the stacking faults evolved into mechanical twins, which released the stress concentration and delayed the necking.

In this paper, the corrosion behavior of three kinds of medium Cr low C steels in the simulated service environment of the transport pipeline was investigated through accelerated corrosion experiments, and the corrosion resistance mechanism of these experimental steels at different temperatures was investigated by electrochemical means. Finally, the reasons for the difference in corrosion behavior were analyzed from the grain boundary and surface Volta potential. The results show that as the temperature rose, the corrosion rate of the 5Cr specimen increased sharply and the corrosion type developed from slight general corrosion to severe general corrosion. The 7Cr specimen was less sensitive to temperature, and the type of corrosion changed from slight general corrosion to local corrosion. Finally, the 9Cr specimen was not sensitive to temperature, and the type of corrosion was always local corrosion. 5Cr steel could form a protective product film at 30°C. As the temperature rose, the protective ability of the product film decreased, and the matrix dissolved easily. The film of 7Cr and 9Cr samples had not yet precipitated and the matrix was difficult to dissolve at 30°C. However, the matrix dissolved easily at 50°C, and the product film had formed, which played a major role. At 70°C, the protective effect of the product film decreased, and the gap between the 7Cr and 9Cr samples began to appear. The increase of Cr content helped to refine grains and increased the proportion of low-angle grain boundaries. At the same time, the increase of Cr element helped to increase the maximum Volta potential of the experimental steel and increased the Volta potential difference. As a result, the test steel was shown to be resistant to uniform corrosion, but it also increased the risk of pitting corrosion.

Event-related potentials (ERP) play an important role in the early detection of emotional arousal. Previous studies of aesthetics have shown that the positive component appearing around 200 ms after stimulus (P2) and the larger late positive component (LPC) are closely related to the early stage of aesthetic judgment. We investigated the temporal features of facial aesthetic judgment on the basis of facial features by using the ERP technique. Participants were instructed to predict holistic face aesthetic level based on the regions of eyes, mouth, or nose. Behavioral results show that holistic score predictions based on the eye region were no different to holistic aesthetic ratings. The ERP analysis results show that beautiful eyes and faces elicited a smaller P2 amplitude and LPC amplitude when judging the holistic aesthetic. The P2 effect of facial aesthetics may reflect automatic processing of facial aesthetics and the difference in LPC may be related to motivational attention to facial aesthetics. Because of the similar ERP effect between the holistic facial aesthetic judgment and the aesthetic judgment of eye region, this region may play a significant role in predicting holistic facial aesthetics. The implication is that the eyes are not only a window to the soul, but also a benchmark of beauty.

There is an urgent need for developing nonprecious metal catalysts to replace Pt-based electrocatalysts for oxygen reduction reaction (ORR) in fuel cells. Atomically dispersed M–Nx/C catalysts have...

In June 2020, an observation experiment of O

We have reported for the first time the synthesis of mixed anatase and rutile TiO2 under 27 MPa in water at the temperature as low as 375 °C. XRD patterns show that the content of rutile in the mixed phase TiO2 is increased by 47.7% compared with the raw TiO2. The average particle size of the mixed phase TiO2 is increased obviously compared with the raw TiO2. Compared with that of the raw TiO2, the band gap of the mixed phase TiO2 is reduced to 2.94 eV while the absorbance is enhanced. Although the particle size increases, the mixed phase TiO2 still exhibits better photocatalytic performance.

Garlic leaf blight caused by

Ridge regression is a method for analyzing multicollinear regression data. In reality, multicollinearity often exists among data. Linear regression based on the principle of ordinary least square method has too large variance due to the unbiasedness of least squares estimation, which leads to it being far from the true value. In essence, ridge regression algorithm is an improved least square estimation method. By giving up the unbiasedness of least square method, the result is more practical, and the problem is indirectly transformed into an optimal parameter problem that minimizes the error function. In the field of machine learning, gradient method is usually used to solve the problem, but it is easy to fall into the local minimum trap. In this paper, a ridge regression based on high order truncated gradient descent method is proposed, and the high order truncated gradient descent method is used to replace gradient descent method.

There is an urgent need for developing nonprecious metal catalysts to replace Pt-based electrocatalysts for oxygen reduction reaction (ORR) in fuel cells. Atomically dispersed M-N

International audience; Inorganic semiconductor heterojunctions have attracted extensive attention for technological applications owing to interface phenomena. In this work, selenide composites with in situ constructed Sb2Se3/NaSbSe2 heterojunctions with similar band gap width have been synthesized by conventional melt-quenching method. The microstructure and photoelectric properties of the composites were investigated at different NaSbSe2 contents. The study revealed that NaSbSe2 content has direct influences on the photoelectric performance of the composites by modifying the heterojunction size and the grain boundary resistance, as evidenced by scanning electron microscopy and impedance spectroscopy, respectively. The photocurrent density of Sb2Se3 with 2.5 and 5 mol% NaSbSe2 was approximately 180 times larger than that of the pure Sb2Se3 at a bias voltage of -1.5 V and under an illumination of 7 mW/cm2, due to an effective separation of photogenerated carriers through the interface electric field. Additionally, the composites exhibited an excellent stability even in 0.5 M LiClO4 aqueous electrolyte.

Relative-risk models are often used to characterize the relationship between survival time and time-dependent covariates. When the covariates are observed, the estimation and asymptotic theory for parameters of interest are available; challenges remain when missingness occurs. A popular approach at hand is to jointly model survival data and longitudinal data. This seems efficient, in making use of more information, but the rigorous theoretical studies have long been ignored. For both additive risk models and relative-risk models, we consider the missing data nonignorable. Under general regularity conditions, we prove asymptotic normality for the nonparametric maximum likelihood estimators.

A low radar cross section (RCS) antenna designed with band-notched absorber is described. First, a dual-polarization absorber with relative bandwidth more than 80% is designed using loaded resistors. Pairs of circular slot resonators and metal strip array resonators are introduced in the absorber to realize a notch band with full reflectance in the vertical polarization, while a wide absorption band in the horizontal polarization is maintained. The proposed band-notched absorber is thus realized. Within the notch band, the absorber can be served as a metal ground for antenna; while a great RCS reduction is obtained out of the notch band and in the horizontal polarization band. Then, a dipole antenna rigorously designed is mounted above the band-notched absorber, whose operating frequency is exactly in accordance with that of the notch band. The proposed low RCS antenna is established based on assembling the dipole antenna and the band-notched absorber together. The measured results demonstrate that the proposed antenna has fairly good radiation patterns. Added to that, more than 10 dB RCS reduction in two polarizations is realized simultaneously compared with that one of a conventional dipole antenna.

A low radar cross section (RCS) and low profile antenna co-designed with absorbent frequency selective radome (AFSR) is investigated. A pair of circular slot resonators is embedded on surface of the AFSR to realize a transmission window in the vertical polarization, while the wide absorption band is still maintained in the horizontal polarization. When a patch antenna is etched within the AFSR, where the metal grounds of the patch antenna and AFSR are co-used, the co-designed antenna with low RCS and low profile is thus realized. For demonstration, an AFSR is designed with its transmission window has a minimal insertion loss of 0.45 dB at 8.9 GHz, and two separate absorption bands (a lower absorption band from 4.8 to 7.5 GHz and an upper absorption band from 10 to 13 GHz) in the vertical polarization, a wide absorption band (from 4.5 to 12.5 GHz) in the horizontal polarization. A patch antenna etched within the AFSR is optimized to operate at 8.9 GHz, then it is simulated and fabricated. The measured results demonstrate that the proposed antenna not only has good radiation patterns, but also obtains significant RCS reduction.

A reproduction and short discussion on the phase and evolution of Li3Fe2(HPO3)3Cl presented in the article entitled Li3Fe2(HPO3)3Cl: an electroactive iron phosphite as a new polyanionic cathode material for Li-ion battery (Hooman Yaghoobnejad Asl, et al., J. Mater. Chem. A, 2015, 3, 7488) are given. According to our reproduction, the major phase in the temperature range by Hooman Yaghoobnejad Asl, et al. is not Li3Fe2(HPO3)3Cl, but is Fe11(HPO3)8(OH)6 and the major impurity phase is not Li1.43[Fe4.43IIFe0.57III(HPO3)6]·1.5H2O as claimed by the authors. More detailed experimental parameters are suggested to be given.

Graphite flexible grounding material is an environmentally friendly material with corrosion resistant, non-magnetic and flexible characteristic, which is gradually applied to grounding of transmission lines in recent years. In order to provide a reference for grounding grid design and impulse resistance measurement, an iterative calculation method based on electric-network theory is proposed. The equivalent method of the equivalent time-varying radius and influence of different iteration methodsonthe calculation resultsare discussed. Furthermore the accuracy of this method is compared with that of traditional ones in regulations. In the end the impulse grounding resistance of three kinds of grounding grids are calculated with this method and the empirical formulas of impulse coefficient are built. The results show that it can be considered that there is reference value in the calculation results obtained by the method in this paper, and the calculation accuracy improved as well. It obtains three typical grid structures under the impulse coefficient experience formula, and the grounding grid design and impulse resistance measurements of flexible graphite grounding electrode can provide reference. Flexible graphite grounding material has a broad application prospect in the power system grounding technology.

A fascinating structure transition of cerium around 5 GPa has been studied using high-resolution synchrotron angle-dispersive powder diffraction techniques. This study indicates that the structural transition of cerium strongly depends on the process of compression. Two runs experiments with different compression conditions demonstrate two different structural transition sequences at room temperature around 5 GPa, fcc′→ C 2/ m →bct for regular compression, fcc′→ C 2/ m →α-U→bct for the pre-compression. These results show that pressure is not the only factor to determining structural transformation at high pressure, and the structural transformation sequences can be varied by changing the external conditions e.g. pre-compression.

In this paper, a band-notched absorber with high notch-band-edge selectivity is proposed and investigated. It is formed by introducing a narrower and full reflectance band into a wider absorption band. For each absorbing unit cell, two independently magnetic and electric resonances with adjacent resonant frequencies are employed to develop the narrower and full reflectance band, while a dipole-shaped metal strip with a resister printed on one side of the supporting substrate and a ground plane is used to realize the wider absorption band. By retrieving relative equivalent circuits of each cell, the performance of the proposed absorber can be qualitatively analyzed. Simply through tuning the frequencies of the independently magnetic and electric resonances, the performance of the notched band can be adjusted with high notch-band-edge selectivity. A sample of an absorber with the notched band frequency from 8.2 to 9.8 GHz, good absorption band from 4.8 to 8.0 GHz and 10.2 to 16 GHz has been designed, fabricated, and measured. Good agreement among circuit analysis, simulation results, and measurement results is finally obtained.

We examined whether or not self-efficacy plays a role in stereotype-activation effect. We found in Study 1 (N = 46) that compared to being primed with the stereotype of an occupation that required less mental effort (cleaner), participants primed with the stereotype of a highly qualified information technology expert performed better in a general knowledge test. Self-efficacy reliably mediated the effect of stereotype activation on test performance. In Study 2 (N = 46) we found that the group primed with the stereotype of an athlete exhibited greater endurance in a physical test than did those who were primed with the stereotype of a homeless person. These results show that self-efficacy beliefs acted as a mediator in the behavior of the participants.

Objective To study the expression of RUNX3 in colorectal adenocarcinoma tissues and its correlation with microvessel density (MVD), and investigate the clinical pathological prognostic significance of RUNX3 and MVD in patients with colorectal cancer. Methods The expression value of RUNX3 and MVD in 70 specimens’ colorectal adenocarcinoma tissues were detected by immunohistochemistry staining technique. The correlation between their expression and the clinicopathologic features was also investigated. Results The expression value of RUNX3 and the positive rates of RUNX3 in colorectal adenocarcinoma tissues were 3.25 ± 1.14 and 25.71% (18/70). The expression value of MVD in colorectal adenocarcinoma tissues was 13.14 ± 3.23. Expression of RUNX3 and MVD value were correlated with CEA, serosal invasion, liver metastasis, lymph node metastasis, and TNM stage (P

Principal component analysis is an unsupervised linear dimension reduction method in machine learning. The basic idea of principal component analysis is to project the data of feature space into low dimensional space, and ensure the maximum variance of projection point, which turns the problem into the optimal parameter problem that can make the maximum variance. In the field of machine learning, the gradient ascent method is usually used to solve the optimal parameters, but it is easy to fall into the local maximum trap. In this paper, a principal component analysis method with high- order truncated gradient ascent method is proposed. High- order truncated gradient ascent method is used to replace the gradient ascent method so as to better retain the gradient information and have better global convergence.

A novel method of developing an electromagnetic absorber is proposed through the design of an ultrawideband resistor-loaded planar dipole absorber. With the method, an antenna array structure is first designed for its best transmission performance and configured as an absorber with the application of the reciprocal principle. The designed planar dipole antenna has achieved a fractional bandwidth of 105% for an absorbing rate of more than 90% under the normal incident. In addition, the thickness of the proposed structure is $0.128\lambda _{0} $ with $\lambda _{0} $ being the wavelength at the center operating frequency. Good agreement between the simulated and measured results demonstrates the validity of the proposed method.

The CO2 gasification properties and kinetics of biomass chars including four kinds of herbaceous residues and two kinds of wooden residues have been studied by the method of isotherm-gravimetric analysis. In addition, the chemical components as well as physical structures of six chars were systematically tested. Results show that gasification reactivity of herbaceous residue char were better than that of wooden residue char. It was found that gasification reactivities of char were mostly determined by its carbonaceous structure. Four kinetic models were applied to describe the gasification behavior of biomass chars: the volumetric model (VM), the grain model (GM), the random pore model (RPM) and the modified random pore model (MRPM). It was found that the RPM and MRPM model were better for describing the reactivity of different chars. However, for the gasification process in which the peak gasification rate appears in high conversion range, the MRPM performs better. At the same time, a marked compensation effect was also presented between the activation energy and pre-exponential factor when the Arrhenius law was used to describe the temperature dependence of gasification rate of char.

CaF2-type UNxOy layers prepared by the nitriding of depleted uranium metal exhibit good oxidation resistance; however, the mechanism is still lacking. This study examined the effect of O on extrinsic point defects in UN2 using density functional calculations. The calculations reveal that N vacancy easily forms in CaF2-type UN2, and O atom lowers the formation energy of N vacancy. By gradually replacing N atoms with O atoms, a series of U–N–O ternary compounds with various N:O ratios, which are not only energetically stable but also mechanically stable within both GGA and GGA+U frameworks, have been found. Additionally, full U–N bonding states contributes more to stabilize the crystal system than the partial U–O bonding states. Collectively, these calculations predict the existence of U–N–O ternary compounds, some of which cannot be distinguished from UN2 by regular X-ray diffraction. Present calculations show that the oxidation resistance of UNxOy layers may be related to U–N–O ternary compounds with various N:O ratios, and suggest further comparative exploration of the adsorption and diffusion of O atoms on the surface of typical ternary compounds, UN2 and UN.

The combustion characteristics of brown coal, anthracite and their blends were investigated by thermogravimetric analysis. The combustion characteristic parameters, including temperature corresponding to ignition, burnout, maximum fractional conversion and mean conversion for the whole process, were analyzed, and the comprehensive combustion characteristic index was calculated as well. A double parallel reactions random pore model (DRPM) was applied to thermal data obtained by non-isothermal thermogravimetric analysis from room temperature to 1173 K at different heating rates in air to calculate the kinetic parameters of anthracite coal–brown coal blends with different ratios. The DRPM provides a good prediction of the fractional conversion curves for co-combustion with coefficient over 0.988. The agreement between experimental and calculated data indicates the accuracy of the DRPM. The activation energy of blends was calculated by DRPM model in the present work as well, and the activation energy of blends decreases with increasing the ratio of brown coal, which show that with increasing the ratio of brown coal in the blends, the combustion reactivity get improved.