Abstract: Code similarity detection is the core component of clone detection.It plays an important role in creating a good learning atmosphere and improving the awareness of intellectual property rights protection.This paper summarizes the processing phases of code similarity detection using fingerprint technique，analyzes the merits and drawbacks of several common methods for fingerprint selection，and proposes an improved method for fingerprint selection orienting code similarity detection.This method is capable of removing false similar ones produced by coding rules from the set of suspected plagiarism code-pairs and obtaining accurate similarity values of code-pairs effectively.

摘要： 代码相似度检测是代码抄袭检测的核心组成部分，对于营造良好的学习氛围、提高知识产权保护意识具有重要作用。总结了应用指纹技术检测代码相似度的基本思路，分析了几种常见的指纹选取方法的优缺点，提出一种改进的面向代码相似度检测的指纹选取方法。该方法能够从疑似抄袭代码集中去除由代码编写规则产生的虚假相似代码对，有效获得代码之间准确的相似度值。