Abstract: In recent years, related studies have put forward methods of obtaining and identifying browser fingerprint. Some website can use dynamic code embedded in the pages to collect browser fingerprint information of the user, including system font setting, display setting, browser version information, etc. However, since the browser fingerprint will be changed when the user upgrades some components of the browser, the method based on static string matching can hardly identify the changed fingerprints. To solve this problem, this paper proposes a new method to identify changed fingerprint based on matching string’s similarity. According to the similarity of the fingerprint, the method can decide whether the collected browser fingerprint is the changed fingerprint or the new one. Based on the proposed method, a website is implemented for collecting and identifying the visitors’ browser fingerprint. The experimental results show that the method can effectively identify the browser’s changed fingerprint so as to improve the recognition rate of returning visit.

Key words: browser fingerprint, static string matching, similarity, changed fingerprint

摘要： 近年来相关研究提出了浏览器指纹的获取和识别方法，网站通过页面中的动态代码来收集用户的系统字体设置、显示设置、浏览器版本等指纹信息。但由于用户的很多行为能导致浏览器的指纹信息产生变化，而完全静态匹配指纹的方法对变化后的指纹无法识别，为了解决该问题，提出了一种基于字符串相似度匹配的面向渐变浏览器指纹的识别方法。该方法能够通过计算指纹信息的相似度来判断是否是用户变化后的指纹，且基于上述识别方法，实现了一种面向渐变浏览器指纹的识别系统，对用户的浏览器指纹进行采集和识别，实验结果验证了该方法能有效地识别出渐变浏览器指纹，提高回访用户的识别率。

关键词: 浏览器指纹, 静态匹配, 相似度, 渐变指纹