RESUMO
Finger-vein biometrics has been extensively investigated for personal verification. Single sample per person (SSPP) finger-vein recognition is one of the open issues in finger-vein recognition. Despite recent advances in deep neural networks for finger-vein recognition, current approaches depend on a large number of training data. However, they lack the robustness of extracting robust and discriminative finger-vein features from a single training image sample. A deep ensemble learning method is proposed to solve the SSPP finger-vein recognition in this article. In the proposed method, multiple feature maps were generated from an input finger-vein image, based on various independent deep learning-based classifiers. A shared learning scheme is investigated among classifiers to improve their feature representation captivity. The learning speed of weak classifiers is also adjusted to achieve the simultaneously best performance. A deep learning model is proposed by an ensemble of all these adjusted classifiers. The proposed method is tested with two public finger vein databases. The result shows that the proposed approach has a distinct advantage over all the other tested popular solutions for the SSPP problem.
RESUMO
Cross-site scripting (XSS) attacks are currently one of the most threatening network attack methods. Effectively detecting and intercepting XSS attacks is an important research topic in the network security field. This manuscript proposes a convolutional neural network based on a modified ResNet block and NiN model (MRBN-CNN) to address this problem. The main innovations of this model are to preprocess the URL according to the syntax and semantic characteristics of XSS attack script encoding, improve the ResNet residual module, extract features from three different angles, and replace the full connection layer in combination with the 1*1 convolution characteristics. Compared with the traditional machine learning and deep learning detection models, it is found that this model has better performance and convergence time. In addition, the proposed method has a detection rate compared to a baseline of approximately 75% of up to 99.23% accuracy, 99.94 precision, and a 98.53% recall value.
RESUMO
Objective: This article aims to study the influencing factors of pgRNA and its change magnitude based on the real world. Methods: A total of 421 patients who were tested for pgRNA were selected. According to the baseline data, the subjects were divided into negative and positive groups. The Chi-square test and logistic regression were used to analyze the influencing factors of pgRNA status. Based on the follow-up data, the rank-sum test and linear regression were used to analyze the influencing factors of pgRNA change magnitude. Results: A total of 153 (36.3%) of the 421 subjects were pgRNA-negative and 268 (63.7%) were pgRNA-positive. Logistic regression analysis showed that positive HBV DNA (OR: 40.51), positive HBeAg (OR: 66.24), tenofovir treatment (OR: 23.47), and entecavir treatment (OR: 14.90) were the independent risk factors for positive pgRNA. Univariate linear regression showed that the pgRNA change magnitude of patients treated with entecavir was higher than that of patients treated with tenofovir. Multivariate linear regression showed that age was an independent factor influencing pgRNA change magnitude. Conclusions: The pgRNA of patients who were young, female, HBV DNA-positive, high-HBsAg, HBeAg-positive is higher than the detection line. HBV DNA and HBeAg are the independent risk factors of positive pgRNA. Different antiviral regimens and disease stages have significantly different effects on pgRNA status. There was a significant correlation between pgRNA and FIB-4, suggesting that pgRNA is related to liver fibrosis. The decrease in pgRNA was greater in young patients than in non-young patients. The decrease in pgRNA was greater in patients treated with tenofovir than in patients treated with entecavir.