基于特征提取和SVM的硬件木马检测方法

针对现有基于机器学习的硬件木马检测方法检测率不高的问题,提出了一种基于特征提取和支持向量机(SVM)的硬件木马检测方法。首先在门级网表的节点中提取6个与硬件木马强相关的特征,并将其作为6维特征向量。然后将这些特征向量分为训练集和测试集。最后使用SVM检测木马。将该方法应用于15个Trust-Hub基准电路,实验结果表明,该方法可实现高达93%的平均硬件木马检测率,部分基准电路的硬件木马检测率达到100%。

Hardware Trojan Detection Method Based on Feature Extraction and SVM

Aiming at the problem of low detection rate of the existing hardware Trojan detection methods based on machine learning, a hardware Trojan detection method based on feature extraction and support vector machine (SVM) was proposed. First, 6 features strongly related to the hardware Trojan were extracted from the nodes in gate-level netlist and taken as 6-dimensional feature vectors for each node. Then, these feature vectors were divided into the training set and the test set. Finally, SVM was used to detect hardware Trojan. The method was applied to 15 Thrust-Hub benchmark circuits. Experimental results showed that this method could achieve an average hardware Trojan detection rate of up to 93%, and the hardware Trojan detection rate of some benchmark circuits could reach 100%.