Search

Glomerulosclerosis is a pathomorphological feature of glomerular lesions. Early detection, accurate judgement and effective prevention of the glomeruli is crucial not only for people with kidney disease, but also for the general population. We proposed a method in combination of traditional image analysis with modern machine learning diagnosis system model based on GoogLeNet for recognizing and distinguishing different categories of glomerulus in order to efficiently capture the important structures as well as to minimize manual effort and supervision. We proposed a novel deep learning model based on GoogLeNet with added batch-normalization layers to extract useful features and subsequently entered the features into SoftMax for classification. We also incorporated Bayesian Optimization algorithm and k-fold cross validation in this system for achieving a more reliable result. Our method has eventually achieved an overall accuracy of 95.04±4.99%, and F1 score of 94.44±3.11% for no glomerulus category, 96.73±5.23% for normal glomerulus category and 93.66±7.82% for globally sclerosed glomerulus category, which means this method can accurately determine the degree of glomerulosclerosis with little supervision. The experimental result also shows that this method has better performance when compared with other state-of art methods.

An efficient and fast method to detect the position and coplanarity of pins is obtained by combining optics and laser in this study. By using the optical camera to position, enlarge and auto focus the ceramic column grid array (CCGA) device, adjusting the light source of lens to realize the transformation of light and dark field, the pin can be accurately identified, and the pin image within the field of view can be collected through the rapid automatic contour capture technology, the pin coordinate can be efficiently extracted, and the device position data can be calculated quickly. Then, the coplanarity calculation would be figured out by scanning the pin height data with a line laser. Through a variety of means such as measuring microscope, the parameters of the detection method and equipment settings are optimized. The coplanarity and position of CCGA devices are detected conveniently, efficiently and quickly, and the detection efficiency and accuracy are significantly improved.

In many cases, the closed-loop phase current control is indispensable for the operation of the switched reluctance machine (SRM). To achieve the compromise among the convenience, adaptability and accuracy, a phase current tracking control approach is proposed, which fully utilizes the robustness of the sliding mode theory. According to the reference phase current from certain algorithm, the reference excitation voltage is obtained with the sliding mode controller, which just requires the unsaturated phase inductance of the SRM at the aligned rotor position. Furthermore, the excitation voltage is accurately supplied to the phase windings through a four-state pulse width modulation (PWM) scheme. The performances of the proposed control method are evaluated with detailed simulations and experiments under different operating conditions, and the comparisons with the commonly used hysteresis current control (HCC) method are carried out as well.

How to achieve hair editing has always been a longstanding topic. The task of hair editing expects users to create a draft to customize hairstyle, textures and colors. Then it will translate the hair draft to corresponding authentic hair. However, hair generally has complex feature representations due to delicate textures and rich colors, which means it difficult to retain all the hair features during the generation process. Thus, how to reconstruct hair with visually pleasant textures and diverse colors is a challenge. To address this problem, this paper proposes a novel method that takes Image-to-Image Generative Adversarial Networks as the backbone and induces two-phase training mode to capture different hair features. In the first phase, networks take the hair colour draft as input to reconstruct low-frequency information. In the next phase, the result output in the first phase will be fed into the network again to refine hair. Our extensive experiments show that two-phase training mode can build up convincing hair images and also allows users to customize the style of hair in the textures as well as colors.

In this paper, a novel three-dimensional (3-D) Frequency Selective Surface (FSS) which is independent of incident angle is proposed. By mapping circular aperture unit elements of two-dimensional (2-D) FSS to spheres, an innovative 3-D FSS which contains annular aperture unit elements is designed. Through controlling structure parameters, the resonant frequency shift characteristics of the proposed 3-D FSS under oblique incidence can exhibit opposite tendency, which provides a possible way to offset the resonant frequency shift of the dielectric slab caused by different incidence angles. The performance of the proposed 3-D FSS loaded with dielectric slab demonstrates excellent resonance stability over a large range of oblique incidence, and the resonance frequency remains constant for different incident angles, even valid up until 80 degree.

The accompanying symptoms of hearing loss is slow and sensory, which makes detecting hearing loss of huge significance to the medical diagnosis and scientific research field. To improve the efficiency of hearing loss classification, we conducted a research on a dataset obtained from magnetic resonance imaging and presented a novel computer aided system based on stationary wavelet entropy, k-fold cross validation, single-hidden-layer feedforward neural network and genetic algorithm. Firstly, the features are extracted from each hearing loss image via stationary wavelet entropy. Then, we used the genetic algorithm to train the single-hidden-layer feedforward neural network. The system reaches an overall sensitivity of 89.89±2.50%, which means the model gives much better performance than manual interpretation.

The vehicle density determines the frequency of information congestion and collision in VANETs, and affects the quality of communication. Cluster management of vehicle nodes can effectively improve communication efficiency of the network. A clustering framework based on changes in vehicle density and an improved k-means clustering algorithm based on vehicle movement characteristics are proposed in this paper. According to the change of vehicle density, nodes are dynamically clustered, in some cases MSCNs(Mobile Secondary Computing Node) are selected and virtual computing areas are divided. The simulation on NS3 shows that the proposed improved k-means-based adaptive clustering algorithm has strong stability and high communication efficiency.

A dual-band frequency selective surface (FSS) with large band ratio (BR) value for 5G communication is proposed and analyzed in this paper. This design is based on Patch-Aperture-Patch (PAP) model, and to cover both low-frequency and high-frequency bands of 5G communication without grating lobes, proper miniaturized elements are employed. In addition, unlike the traditional PAP structure, the exterior patches of designed unit don’t share the same dimension with the purpose of increasing the BR value. Simulation results show that the proposed FSS structure can generate band-pass response at 5 and 27 GHz, which are exactly the operation bands of 5G communication. In the meanwhile, there is no transmission window observed between two pass-bands and the out-of-band suppression is below -10 dB. The designed FSS has the advantages of low insertion loss, large band ratio, excellent polarization and azimuth stability and high out-of-band suppression, making it suitable for multiband 5G communication in the future.

This article designs a slotted frequency selective surface (FSS) and obtains effectively microstrip antenna radar cross section (RCS) reduction based on this novel FSS. The characteristics of the slotted FSS are analyzed in contrast with the normal FSS. The reference antenna is a microstrip antenna with normal FSS ground, and the modified antenna is a microstrip antenna with slotted FSS ground. The radiation performance and RCS between two antennas are compared on the basis of simulation analysis. As a result, the modified antenna with slotted FSS ground realize favorable RCS in-band and out-of-band reduction in contrast with reference antenna. And 12 dBsm maximum RCS reduction can be achieved at the frequency of 3.05 GHz.

Integrating solar PV energy into marine power system provides a new way to deal with the rapid growth of fossil fuels consumed by traditional ships in improving the energy efficiency and reducing the greenhouse-gas emissions. In this study, one practical PV/SC hybrid power system integrated into a 2240 pure car and truck carrier ro-ro ship is utilized as a case to carry out the techno-economic analysis. Taking into account the navigation route of the target ship, the amount of electricity generated by the PV/SC hybrid system is calculated based on the solar irradiation data of the seven port cities on the route. The analysis results show that (i) the levelized cost of energy of the PV/SC hybrid power system reduces to 0.53 USD/kWh from 3.89 USD/kWh over the life cycle; (ii) the discounted payback time of the PV/SC hybrid power system is 5.49 year. 1

This paper mainly studies the electromagnetic immunity of two kinds of static random access memory (SRAM) chips using bulk silicon (SI) and insulator silicon (SOI) technologies respectively, and the effect of temperature on chip immunity. Direct power injection (DPI) method was used to test the immunity of each functional pin of the two chips, and the ground pin was selected to study the immunity under heat stress. The test results show that the immunity of each functional pin is different, the immunity of the same pin is also different under different working states, and the SOI technologies has better immunity than the SI technologies. At the same time, it was found that different temperatures would cause the chip’s immunity threshold to drift. This provides an important reference for the design of the chip immunity.

With the development of the CPU industry, the electromagnetic compatibility (EMC) of CPU and its testing methods are receiving more and more attention from electronics, electrical engineers and engineering technicians. In order to meet the requirements of electromagnetic compatibility standard testing and to ensure the stability and reliability of the CPU, the designer must perform complete electromagnetic compatibility testing and design of the product. In this paper, the CPU chip is tested in the near field scan (NFS) according to the IEC61967–3 standard. Compared with the traditional test, the surface scanner is used for testing. This paper analyzes the test method, analyzes the test results in detail, lays a foundation for in-depth study of the electromagnetic compatibility of the CPU, and provides a basis and reference for the domestic IC EMC standard formulation. In addition, this paper also puts forward some considerations in the actual testing process of the surface scan method.