VPP_AHA: Visual Privacy Protection via Adaptive Histogram Adjustment

Multimedia Tools and Applications - A novel visual privacy protection (VPP) algorithm is proposed for camouflage of the visual identifiers in digital images. The challenge is to suffice the...     

Mathematical law of size effect on the flexural property of ceramics

Brittle materials generally exhibit size effects, and the mechanical properties of these materials degrade significantly with an increase in size. However, the mathematical law governing the attenuation degree of mechanical properties with the increase in size is still unknown. In this study, maximum loads of differently sized ceramic test strips were subjected to three point bending tests under two working conditions of equal spans and span amplifications, respectively. Subsequently, the theoretical maximum loads of materials were calculated using the finite element method (FEM). By calculating the difference between the calculated values and the actual maximum loads, the attenuation of mechanical properties of ceramic samples were observed. The results show that the theoretical mechanical properties and the performance attenuation caused by the size effect tend to increase according to the following equation: y=ax³+bx²+cx+d. Therefore, mechanical properties and performance attenuation of any sample exhibiting a size within the experimental range can be predicted by a mathematical law, which was obtained through mechanical tests results of four samples with different sizes. The obtained mathematical law holds great significance for predicting the mechanical properties of materials under size effects.     

Synthesis of cubic and monoclinic hafnia nanoparticles by pulsed plasma in liquid method

Hafnia (HfO₂, hafnium dioxide) is a wide band gap and high-κ material, and the metastable cubic hafnia has a much higher permittivity compared with the normal monoclinic hafnia. Here, we employ a one-step process, the pulsed plasma in liquid (PPL) method to synthesize two types of hafnia nanoparticles (NPs): one which is mainly in cubic phase (cubic: 81.7 at%, monoclinic: 18.3 at%) and the other which is in monoclinic phase. High-resolution transmission electron microscopy images showed that the particles were small (particle size ~3 nm). X-ray absorption fine structure analysis showed no chemical shifts, indicating that the synthetic hafnia NPs contained no oxygen vacancy. The synthetic hafnia NPs mainly in cubic phase showed a much higher relative permittivity than that of the commercial hafnia (monoclinic), and have a larger band gap than the synthetic monoclinic hafnia NPs.     

Self-terminated electrodeposition of platinum on titanium nitride for methanol oxidation reaction in acidic electrolyte

The development of high-performance electrocatalysts for methanol oxidation is an urgent task to enhance the efficiency of direct methanol fuel cells. We report a simple and controllable method to fabricate Pt-decorated TiN electrocatalysts using self-terminated electrodeposition at room temperature and ambient pressure. Under optimized deposition parameters such as electrolyte pH, TiN substrate pretreatment, and pulsed deposition potential, quenching of the Pt electrodeposition facilitates obtaining an extremely low Pt mass loading (0.93 μg/cm²) on the TiN substrate. Repeated deposition potential pulses enable a gradual increase in Pt loading, with a precise control of the loaded Pt mass. Maximum intrinsic and mass activities for the methanol oxidation reaction are achieved for the catalyst with a Pt loading mass of 55.0 μg/cm², prepared by 20 deposition pulses. The maximum intrinsic activity achieved with the Pt-decorated TiN electrocatalyst is five times higher than that obtained with bulk Pt. The present results thus provide a facile method for the fabrication of cost-effective electrocatalysts.     

Characterization of the ash deposits along deep-cooling of the exhaust gas in coal-fired power plants

Ash deposition and condensation of acid vapors under deep-cooling of exhaust gas in a 1000 MW ultra-supercritical power plant were investigated. X-ray diffraction, X-ray fluorescence and laser particle size analyzers were employed to characterize the components, elements and particle size distribution of deposits at different metal wall temperatures. The results show that the deposits formed at the metal wall temperatures of 40 °C–60 °C were much thicker than the deposits at 70 °C–90 °C from the perspectives of the morphology and thickness. Fluorides and chlorides were observed in deposits formed at the metal wall temperatures of 40 °C–60 °C. Thin deposits were composed of Al-Si oxides and simple sulfates when the metal wall temperatures were 70 °C–90 °C. The dew-point temperatures of water, HCl and HF vapors were around 40 °C, 60 °C, and 60 °C, respectively. Basic ferric, aluminum and calcium sulfates, which were acted as traps to absorb the condensate and fly ash. Hydration played an important role in the growth of deposits formed at the metal wall temperature of 40 °C–60 °C.     

Applying Feature-Weighted Gradient Decent K-Nearest Neighbor to Select Promising Projects for Scientific Funding

Due to its outstanding ability in processing large quantity and high-dimensional data, machine learning models have been used in many cases, such as pattern recognition, classification, spam filtering, data mining and forecasting. As an outstanding machine learning algorithm, K-Nearest Neighbor (KNN) has been widely used in different situations, yet in selecting qualified applicants for winning a funding is almost new. The major problem lies in how to accurately determine the importance of attributes. In this paper, we propose a Feature-weighted Gradient Decent K-Nearest Neighbor (FGDKNN) method to classify funding applicants in to two types: approved ones or not approved ones. The FGDKNN is based on a gradient decent learning algorithm to update weight. It updatesthe weight of labels by minimizing error ratio iteratively, so that the importance of attributes can be described better. We investigate the performance of FGDKNN with Beijing Innofund. The results show that FGDKNN performs about 23%, 20%, 18%, 15% better than KNN, SVM, DT and ANN, respectively. Moreover, the FGDKNN has fast convergence time under different training scales, and has good performance under different settings.     

Interfacially reinforced methylphenylsilicone resin composites by chemically grafting multiwall carbon nanotubes onto carbon fibers

Both silane and multiwall carbon nanotubes (CNTs) were grafted successfully onto carbon fibers (CFs) to enhance the interfacial strength of CFs reinforced methylphenylsilicone resin (MPSR) composites. The microstructure, interfacial properties, impact toughness and heat resistance of CFs before and after modification were investigated. Experimental results revealed that CNTs were grafted uniformly onto CFs using 3-aminopropyltriethoxysilane (APS) as the bridging agent. The wettability and surface energy of the obtained hybrid fiber (CF-APS-CNT) were increased obviously in comparison with those of the untreated-CF. The CF-APS-CNT composites showed simultaneously remarkable enhancement in interlaminar shear strength (ILSS) and impact toughness. Moreover, the interfacial reinforcing and toughening mechanisms were also discussed. In addition, Thermogravimetric analysis and thermal oxygen aging experiments indicated a remarkable improvement in the thermal stability and heat oxidation resistance of composites by the introduction of APS and CNTs. We believe the facile and effective method may provide a novel interface design strategy for developing multifunctional fibers.     

Novel DDoS Feature Representation Model Combining Deep Belief Network and Canonical Correlation Analysis

Distributed denial of service (DDoS) attacks launch more and more frequently and are more destructive. Feature representation as an important part of DDoS defense technology directly affects the efficiency of defense. Most DDoS feature extraction methods cannot fully utilize the information of the original data, resulting in the extracted features losing useful features. In this paper, a DDoS feature representation method based on deep belief network (DBN) is proposed. We quantify the original data by the size of the network flows, the distribution of IP addresses and ports, and the diversity of packet sizes of different protocols and train the DBN in an unsupervised manner by these quantified values. Two feedforward neural networks (FFNN) are initialized by the trained deep belief network, and one of the feedforward neural networks continues to be trained in a supervised manner. The canonical correlation analysis (CCA) method is used to fuse the features extracted by two feedforward neural networks per layer. Experiments show that compared with other methods, the proposed method can extract better features.     

R-curves determination of ordinary refractory ceramics assisted by digital image correlation method

As a figure-of-merit, the rising ratio of crack propagation resistance to fracture initiation resistance indicates a reduction of the brittleness and enhances the thermal shock resistance of ordinary refractory ceramics. The significant nonlinear fracture behaviour is related to the development of a fracture process zone (FPZ). The universal dimensionless load–displacement diagram method is applied as a promising graphical method for the determination of R-curves for magnesia refractories showing different brittleness. By applying digital image correlation (DIC) together with the graphical method, the problems arisen with accurate determination of the fracture initiation resistance and the crack length are overcome. Meanwhile, the R-curve is subdivided with respect to the fracture processes, viz the fracture initiation, the development of FPZ and the onset of traction free macro-crack. With the simultaneous crack lengths evaluated from DIC, the contribution of each fracture process to the crack propagation resistance at certain loading stage is quantitatively presented.