摘要
为了快速精确地检测识别恶意代码,发展其检测关键技术显得尤为必要。本文在恶意代码可视化的理论基础上提出一种新的基于纹理特征的恶意代码检测方法,通过可视化算法实现对恶意代码二进制程序的可视化显示,即映射为灰度图像,再提取图像的纹理特征,并在恶意代码语料库中进行特征匹配,最后输出检测结果。利用在某公司安全部门捕获到的5种出现次数最多的恶意代码种类进行实际检测实验,实验结果表明基于纹理特征的恶意代码检测方法实用性强,对恶意代码的分类非常精确。
It is necessary to develop the key technology so as to detect and identify malicious code quickly and accurately. A new texture-feature-based approach is proposed to extract and detect the feature from malware content, which is based on the theory of malicious code visualization. The malicious code is mapped to uncompressed gray-scale image by the visualization of binary program of the malicious code; the texture-feature is extracted by gray-scale B2V algorithm; then matches the texture-feature in the corpus of malicious code; finally output the matching results. At last, using 5 types of malicious code captured on Yunnan Power Network Information Center website, which are appeared most frequently, to make a practical experiment. The experimental results show that this detection method has strong practicability, and the classi？cation of malicious code by using this method is very accurate.
作者
汪应龙
黄祖源
刘爱莲
李川
Wang Yinglong;Huang Zuyuan;Liu Ailian;Li Chuan(Automation in Kunming University of Science and Technology, Kunming 650500, China;Information Centre, Yunnan Power Grid Limited Liability Corporation, Kunming 650217, China.)
出处
《云南电力技术》
2018年第1期131-133,138,共4页
Yunnan Electric Power
关键词
恶意代码检测
可视化算法
纹理特征
特征匹配
malware detection
visualization algorithm
texture-feature
feature matching
