中文摘要：

现有协议格式提取方法在语法层次对程序执行轨迹进行分析，字段识别结果可能存在冗余和冲突。为了提高字段识别准确率，提出了一种语义层次的协议格式提取方法。方法首先将执行轨迹中的二进制指令转换为语义等价的中间语言形式，并通过细粒度的动态污点分析跟踪字段语义解析过程，在此基础上，依据字段的语义不可分割性，利用语义层次的字段识别策略实现了协议格式提取。测试结果表明，该方法具有较高的识别精度和较低的分析复杂度。

英文摘要：

Present methods for protocol format extraction analyze the execution traces of programs at syntax level, which leads to redundancy and conflict in the results of field identification. In order to improve the accuracy of field identifica tion, a semantic level method was proposed for protocol format extraction. The method firstly translated the binary in structions into equivalent intermediate language, and then tracked the parsing process of field semantics through finegrained dynamic taint analysis. Further, it extracted protocol format using semantic level policies of field identifica tion, based on the semantic indivisibility of fields. Experimental results show that the proposed method can achieve high identification accuracy with low complexity.