融合DRAE与SVM的网页防篡改检测

针对传统的网络安全研究如入侵检测、流量分析以及主动防御等方法需要较强的网络安全相关知识以及大量的网络训练数据，以及较高的研究门槛的问题，该研究提出一种基于Deep Residual Auto-Encoder（DRAE）与支持向量机（SVM）相结合的网页防篡改检测模型，该模型用DRAE提取网页图像特征，并输入SVM分类器判别网页是否被篡改。经过在东北农业大学范围内实验验证，结果表明，使用该模型进行网页检测的准确率高达95%，高于现有检测方法。

Fusing DRAE and SVM for webpage tamper-resistant detection

Traditional cyber security research methods such as intrusion detection, traffic analysis, and active defense require strong network security related knowledge, a large amount of network training data, and a high research threshold. This study proposes a webpage tamper-proof detection model based on Deep Residual Auto-Encoder (DRAE) and support vector machine (SVM). The model uses DRAE to extract webpage image features and input SVM classifier to determine whether the webpage is tampered with. After experimental verification in Northeast Agricultural University, the results show that the accuracy of web page detection using this model is as high as 95 %, which is higher than the existing detection methods.