Search

In recent years, phishing events occur frequently, and the detection of phishing URL has become a common concern in the field of network security. In previous studies, researchers distinguish phishing URLs from normal URLs by the string characteristics. However, this method is difficult to achieve efficient detection accuracy in the case of attacker's forgery of URL features. In this paper, we propose a phishing URL detection method based on URL content. Firstly, the simple features are used for preliminary screening, and then, the URLs that cannot be distinguished by simple features are input into the simulation environment to obtain the page content. Finally, the method based on CNN is used to detect malicious URLs based on the page content. Experiments show that our method can achieve satisfying detection accuracy.

Polarization lidar has been widely used in recent decades to observe the vertical structures of aerosols and clouds in the atmosphere. To obtain more information from polarization lidar measurements, we developed a dual-polarization lidar system that can detect polarization measurements simultaneously at both 355 nm and 532 nm. The vertical distributions of atmospheric aerosols and clouds over northern China were successfully observed by the developed lidar. Observational data during two typical cases (dust events and haze episodes) were used for the analysis in this study. The results showed that for dust-dominated aerosols, the depolarization ratio (DR) at 532 nm was larger than that at 355 nm, but that for air pollutants was smaller. Our results also show that dual-polarization measurements can be used to largely improve aerosol classification. Moreover, we found that there is a good relationship between the absorption coefficient of aerosols and the ratio of DRs at 532 nm and 355 nm for dust aerosols. These results confirm that the absorption characteristics of dust aerosols cause a difference in DR at the UV and VIS wavelengths, and implying that aerosol absorption may be determined by polarization lidar at the ultraviolet and visible wavelengths.

Accurate identification of aerosols and cloud from remote sensing observations is of importance for quantitatively evaluating their radiative forcing and related impacts. Even though polarization lidar has exhibited a unique advantage of classifying atmospheric aerosols and clouds over the past several decades, polarization measurements are often achieved at one wavelength (UV or VIS) using laser remote sensing. To better identify the types of aerosols and clouds, we developed a ground-based dual-polarization lidar system that can simultaneously detect polarization measurements at wavelengths of 355 nm and 532 nm. Our results show that the volume depolarization ratios (VDRs) at 355 nm and 532 nm markedly differ for typical types of aerosols and clouds in the atmosphere. For non-spherical particles, the ratio of VDRs at 532 nm and 355 nm are 2.87 ± 1.35 for ice cloud and 1.51 ± 0.29 for dust-dominated aerosols, respectively. However, for spherical particles, the ratios are 0.43 ± 0.26 for water cloud and 0.56 ± 0.05 for air pollutants. Consequently, we proposed a simple reliable method for classifying atmospheric aerosols and clouds from polarization measurements observed by the developed lidar system. The proposed method first distinguishes clouds from aerosols using a combination of the color ratio (CR, 532 nm/355 nm) and attenuated backscattering coefficients (ABC) at 532 nm. Then, subtypes of clouds and aerosols are identified based on the ratio of VDRs at 532 nm and 355 nm. The results showed that dual-polarization lidar measurements can remarkably improve the classification of atmospheric aerosols and clouds, compared with results using a traditional method. This study illustrates that more information on atmospheric aerosols and clouds can be obtained from polarization measurements at multiple wavelengths by active remote sensing.

Summary: Background/Objective: The selection of surgical technique in patients with cervical spondylotic myelopathy relies on the surgeon(s) and patients' conditions. The objectives of the study were to test the hypotheses that French-door laminoplasty recovers faster than laminectomy and has good outcome measures. Methods: Data regarding surgical, radiological, and clinical outcome measures of 330 patients with cervical spondylotic myelopathy operated under French-door laminoplasty (fdLP group, n = 110), open-door laminoplasty (odLP group, n = 110), or laminectomy (LC group, n = 110) were collected from the records of institute and analyzed. Results: Patients of fdLP group (p

The interlayers, including hole transporting layer (HTL) and electron transporting layer (ETL), segregating photoactive layer and the electrodes play an important role in charge extraction and transportation in perovskite solar cells (pero-SCs). Two novel copolymers, PDTSTTz and PDTSTTz-4, for the first time were applied as HTL in the n-i-p type pero-SCs, with the device structure of ITO/compact TiO2/CH3NH3PbI3-xClx/HTL/MoO3/Ag. The highest occupied molecular orbitals (HOMO) levels of PDTSTTz and PDTSTTz-4 exhibit a suitable band alignment with the valence band edge of the perovskite. Both of them lead to improved device performances compared with reference pero-SCs based on P3HT as HTL. To further balance the charge extraction and the diffusion length of charge carriers, pristine C60 was introduced at the cathode side of the pero-SCs, working together with TiO2 as ETL. With insertion of both the HTL and ETL, the performance of pero-SCs was greatly enhanced. The optimized devices exhibited impressive PCEs of 14.4% and 15.8% for devices based on PDTSTTz and PDTSTTz-4. The improved performance is attributed to better light harvest ability, decreased interface resistance and faster decay time due to the introduction of the interlayers.

Skin cancer is one of the most lethal cancers in the world, which is a common type of cancer. And the early diagnosis of it is quite important. Aiming to improve the classification of skin cancer, a modified convolutional neural net- work model is proposed. Firstly, a convolutional layer and a pooling layer is added on the original network, constructed an 8-layer convolutional neural network. And then the network output value and the data label by the backward propagation adjust the network parameters. Lastly, by constantly debugging the network, the best model for the recognition network is selected. The proposed network applied to the dataset consisting of 900 cases of melanoma with the malignancy status pathologically confirmed. The results of the revised network prediction accuracy got 91.92% on training set and 89.5% on test set. The results obtained demonstrate that the proposed method is performant and can be used to predict whether the skin cancer is benign or malignant. With the development of deep learning and the clinical application of medical images, this method is expected to become a diagnostic tool for doctors to diagnose skin cancer.

In this paper, triple cathode buffer layers (CBLs) composed of phenyl-C61-butyric acid methyl ester (PCBM), C60, and LiF layers were introduced into the planar p-i-n perovskite solar cells (p-i-n PSCs) with a device structure of ITO/PEDOT:PSS/CH3NH3PbI3-xClx/CBLs/Al. For comparison, a single CBL of PCBM and a double CBL of PCBM/LiF were also investigated in the p-i-n PSCs. On the basis of the PCBM buffer layer, the addition of a thin LiF layer facilitated the charge collection process and led to the dramatic improvement of the power conversion efficiency (PCE) of the PSCs up to 14.69% under an illumination of AM 1.5G, 100 mW/cm(2), which is to date one of the highest efficiencies of the p-i-n PSCs. By further insertion of a C60 layer between PCBM and LiF in the triple CBLs, a PCE of 14.24% was obtained, and more importantly, the PCBM/C60/LiF triple CBLs are very helpful for improving the stability of the devices and making the LiF layer less thickness-sensitive for achieving high performances of the p-i-n PSCs.

Background and Objectives: The diagnosis of cancer is concerned, and the prediction of cell carcinoma is of great importance for the treatment. Materials and Methods: First, we obtain a series of slices of tumor cell pathology in clinical data, with being followed training sets and test sets gained by adding data model. Then, we design a convolutional neural network training and prediction model. After that, we optimize parameters for training and prediction model, combining experience. Results: In experiment, the accuracy of the model predicting for cell carcinoma is 87.38%. Conclusions: This study provides a reference that predicts the extent of cell carcinoma progression by using deep learning model.