浏览器指纹技术的研究与应用-《计算机技术与发展》

文章信息/Info

Author(s):: YANG Li-peng; WANG Tuo; FAN Chun-mei; China Academy of Railway Sciences,Beijing 100081,China

Keywords:: browser fingerprint; device fingerprint; JavaScript; Web security

摘要:: 为了防止用户的信息被盗用,能够及时地发现用户的不正常登录,可以通过研究用户使用的浏览器特征,建立唯一标识用户的指纹信息。由于基于信息熵的浏览器指纹计算算法,会将发生细微变化的同一指纹生成不同的指纹,因此增加了指纹识别的错误率。为了改进该方法,介绍了基于浏览器指纹信息熵和相似度计算的浏览器指纹识别算法,即将不同的浏览器特征通过信息熵计算公式得到一个 ID,当出现同一用户而信息熵不一致时,根据相似度的计算进行判断。该算法综合考虑了指纹完全匹配和部分匹配的情况,提高了同一指纹的辨别能力。实验结果表明,该算法可以更好地识别用户。

Abstract:: In order to prevent the user’s information from being stolen and find the abnormal login of users,we can study the user’s browser
features and establish an unique ID that identifies the user’s fingerprint. As the browser fingerprint algorithm based on the information entro-
py will generate a different fingerprint for the same one with subtle changes,the error rate of fingerprint recognition is increased. In order to
improve above method,we introduce the browser fingerprint identification algorithm based on the information entropy of the browser finger-
print and the similarity calculation. It gets an ID by information entropy calculation formula for different browser features. When the same us-
ers occur but their information entropy don’t match,it is determined according to the similarity of the calculation. The proposed algorithm
comprehensively considers the perfect matching and partial matching for fingerprint,which improves the discrimination to the same finger-
print. The experiment shows that it can be better identification of the use