Xinjiang Police College

In view of the limitations of traditional deep convolutional neural network models in malware classification, such as low accuracy and low precision, this paper proposes an image malware classification method based on AlexNet transfer learning and MHSA attention mechanism. AlexNet is used as the base network, and ImageNet pre-trained image classification model is initialized. By freezing the underlying weights, fine-tuning the output of the last fully connected layer, and introducing transfer learning as training parameters, the overfitting of the model is prevented. Add an MHSA module to the 10th layer (the 5th convolutional layer) of features to capture long-distance dependencies in malware images. Experiments are conducted on the public malware classification dataset maleVis. Compare the information such as accuracy rate, 2-D dimensionality reduction visualization of semantic features in t-SNE, and image regions of interest in Grad-CAM heat map. The results show that the highest classification accuracy of this method on the dataset maleVis reaches 87.3%, which can effectively improve the classification accuracy and precision of malicious software, verifying its feasibility and effectiveness.

Script identification is the first step in most multilingual text-processing systems. To improve the time efficiency of script identification algorithms, whether there is content written in a certain script in the text is first determined; if so, the content written in that script is then obtained. Then, it is determined whether the total length of the texts corresponding to the identified scripts is equal to the original text length; if so, the script identification process ends. Finally, considering the frequencies of various scripts on the Internet, those that are more common are prioritized during script identification. Based on these three approaches, an improved script identification algorithm was designed. A comparison experiment was conducted using sentence-level text corpora in 263 languages written in 26 scripts. The testing times of the newly proposed method were reduced by 9.35-fold, while the F1 score for script identification was slightly higher than those reported in our earlier studies. The method proposed in this study effectively improves the time efficiency of script identification algorithms.

Cross-chain data sharing in decentralized storage networks faces challenges in security and performance, including data breaches and high latency. This paper proposes a layered protocol integrating decentralized storage and cross-chain communication to address these issues. The protocol ensures robust security and optimizes performance through sharding, encryption, and concurrency. Experimental results show significant improvements in throughput and latency compared to baseline schemes, with manageable storage overhead. This study offers a secure, efficient solution for cross-chain data sharing, with applications in decentralized finance and distributed computing. Future work will focus on high-load optimization and real-world validation.

With the rapid development of satellite technology, satellite image processing technology is gradually mature, which can provide reliable data support for various fields. However, in satellite images, vehicle type targets face the problems of complex background, small target size and low contrast, which seriously affect the effect of subsequent experiments. In this paper, an image preprocessing and image enhancement method is proposed to improve the visibility of vehicle targets, reduce background interference, and provide support for subsequent vehicle detection and tracking tasks. The experimental results show that the proposed method can significantly improve the image processing effect and has high robustness.

Aiming at the problem of motor vehicle drivers' fatigue driving, this paper proposes a deep learning-based fatigue driving behavior recognition method. The four parameters of PERCLOS parameter, blinking frequency, continuous eye closure time and continuous mouth opening time are used as the recognition indexes to establish a driver fatigue driving recognition model based on deep learning, and the experimental results show that the method proposed in this paper has a better recognition effect in different light environments.

Urban road traffic safety has become a social problem related to people's life safety, the main factors affecting the occurrence and severity of road traffic accidents are human factors, time, month, vehicle conditions, lighting, road conditions, weather conditions, visibility, etc. Using descriptive and predictive modeling, descriptive modeling is based on Hive database software to analyze the full sample data or Pyechart to visualize the data, the model studies and charts on historical data to obtain corresponding data regularity. Predictive models use machine learning algorithms, accident severity is taken as the target variable, through data cleaning, PCA was used for dimension reduction, RobustScaler data normalization and other processing, select the feature variables, Random forest algorithm or Xgboost algorithm was used to analyze the importance of characteristic variables that affect the severity of accidents. So as to find out the rules, in order to take preventive measures, reduce traffic accidents, to ensure people's safety.

One of the major reasons for the delayed wound healing in diabetes is the dysfunction of endothelial progenitor cells (EPCs) induced by hyperglycaemia. Improvement of EPC function may be a potential strategy for accelerating wound healing in diabetes. Procyanidin B2 (PCB2) is one of the major components of procyanidins, which exhibits a variety of potent pharmacological activities. However, the effects of PCB2 on EPC function and diabetic wound repair remain elusive. We evaluated the protective effects of PCB2 in EPCs with high glucose (HG) treatment and in a diabetic wound healing model. EPCs derived from human umbilical cord blood were treated with HG. The results showed that PCB2 significantly preserved the angiogenic function, survival and migration abilities of EPCs with HG treatment, and attenuated HG‐induced oxidative stress of EPCs by scavenging excessive reactive oxygen species (ROS). A mechanistic study found the protective role of PCB2 is dependent on activating nuclear factor erythroid 2‐related factor 2 (Nrf2). PCB2 increased the expression of Nrf2 and its downstream antioxidant genes to attenuate the oxidative stress induced by HG in EPCs, which were abolished by knockdown of Nrf2 expression. An in vivo study showed that intraperitoneal administration of PCB2 promoted wound healing and angiogenesis in diabetic mice, which was accompanied by a significant reduction in ROS level and an increase in circulating EPC number. Taken together, our results indicate that PCB2 treatment accelerates wound healing and increases angiogenesis in diabetic mice, which may be mediated by improving the mobilization and function of EPCs.

SummaryAmong the commercial materials, stainless steel is widely used in our daily life and can be hardly destroyed by flame or the heat emanating from a fire. In this fundamental work, the formation and development of oxide scale on stainless steel 1Cr11Ni2W2MoV were investigated at 600°C to 800°C in the atmospheres with and without liquid accelerant. The aim of this work is to figure out the influence of accelerant on the oxidation pattern of stainless steel at high temperature that occurs in a fire. The morphology, microstructure, and the growth rate of the oxide scales were characterized by thermogravimetric analysis, visual analysis, scanning electron microscopy, energy‐dispersive spectroscopy, and X‐ray diffraction. The results revealed that the oxide scale formed on the stainless steel 1Cr11Ni2W2MoV was mostly protective in both atmospheres from 600°C to 800°C, except that breakaway oxidation occurred locally that resulted in the formation of intrusion oxide. Both increasing temperature and the presence of kerosene in combustion atmosphere increased the mass gain of stainless steel, which was mainly attributed to the occurrence of local breakaway oxidation. Consequently, the addition of accelerant just increased the formation trend of local intrusion oxide, rather than remarkably affect the scaling behavior. Therefore, careful analysis is needed to identify the presence possibility of accelerant in oxidation atmosphere according to the scaling behavior of stainless steel. Characterization of surface scale and metallurgical analysis of metallic material are expected to be supplementary technique for fire characterization in the future.