基于感知哈希算法的Android恶意软件入侵检测仿真

Android malware intrusion detection simulation based on perceptual hash algorithm

为保护智能终端用户隐私信息安全、避免恶意软件入侵破坏终端功能,提出一种基于感知哈希算法的Android恶意软件入侵检测方法。令Watson感知模型和哈希算法结合,将恶意软件代码可视化变换为灰度图像,经离散余弦变换,得出代码特征对比度掩蔽表达矩阵,使用明科斯基法获得感知哈希序列距离误差。考虑到同家族软件代码相似度高,为避免恶意开发人员通过伪装或者混淆代码方式使检测方法失效,计算代码局部特征的权重值,明确不同特征对哈希序列相似度检测结果的影响,最后与历史数据库内已知的恶意软件代码对比感知哈希序列,输出检测结果。经过仿真证明,所提方法能够精准检测出入侵的恶意软件,且误报率低,检测耗时短,具有极高的应用价值。

In order to protect the privacy information of intelligent terminal users and prevent malicious software intrusion from destroying terminal functions,this article presented a method of detecting malware intrusion in Android based on perceptual hash algorithm.Firstly,Watson perception model was combined with hash algorithm,the malware code was visualized as a gray image.After the mask matrix of code feature contrast was obtained through discrete cosine transform.Meanwhile,the distance error of perceptual hash sequence was obtained by Kominsky method.Due to high similarity of the same family of software codes,in order to prevent malicious developers from invalidating the detection method through camouflage or code confusion,the weight value of the local features of code was calculated to clarify the impact of different features on the hash sequence similarity detection results.Finally,the perceptual hash sequence was compared with the known malicious software codes in historical database,thus outputting the detection results.The simulation results prove that the proposed method can detect the malicious software intrusion accurately,with low false alarm rate and short detection time,so it has high application value.