基于函数调用图的Android恶意软件检测

Android Malware Detection Based on Function Call Graph

为了有效检测Android恶意软件,提出一种基于函数调用图的检测方法。通过将应用程序的函数调用关系表示为图结构,结合深度学习处理图的算法对图结构经过节点排序、归一化等处理,生成能够输入卷积神经网络的局部感受野从而建立恶意软件分类模型。通过和不同的恶意软件检测模型对比,证明基于图结构的检测方法有较高的检测准确率和检测效率。

In order to detect Android malware effectively, proposes a method based on function call graph. By expressing the function call relation of the Android application as the graph structure, uses the algorithm of the deep learning for graphs to process the graph structure through node ordering, normalization and so on, and generates the local receptive field of the convolutional neural network to establish the malware classification model. The Experiments show that the detection method has a high detection accuracy and efficiency by comparing with dif- ferent malware detection models.