基于同态加密的密文检索方案研究

Cipher text Retrieval Technology Based on Homomorphic Encryption

摘要:

现有的密文检索技术主要是采用的是布尔模型,它无法精确的计算出检索项与待检索文件的相关度,不能按相似度进行精确的排序；针对以上情况,结合同态加密技术和基于TF-IDF的向量空间模型技术,提出了一个基于向量空间模型全同态环境下的密文检索方案BVH(based vector space model and homomorphism ciphertext retrieval scheme),BVH主要分为3个步骤:第一是预处理阶段,主要对上传的文件建立倒排索引,生成文件向量集,计算各个文件向量的模,对文件向量集和要上传的文件加密以密文的形式上传到云端；第二个阶段是检索阶段,主要是将搜索词的向量密文和各个文件向量的密文相乘将结果以密文的形式返回给客户端；第3个阶段结果处理阶段,主要是对解密后的结果进行相应的计算处理,对最后的处理结果按相似度大小排序；经分析,该方案在准确率和检索效率方面都得到了较大提升。

Abstract:

The existing cipher text retrieval techniques mainly adopt the Boolean model and can not calculate accurately the related degree of the search terms and the retrieved documents, then it can not sort precisely. In view of the above situation, combined with homomorphic encryption technology and vector space model technique based on TF-IDF, the paper proposed a cipher text retrieval scheme BVH , BVH is divided into three steps:the first is preprocessing stage, it's main work is the establishment of inverted index file and vector set, and the calculation of each document vector module, then uploads the encrypted document vectors and files to the cloud. The second stage is the retrieval stage, it mainly multiplys key words vector cipher text and each file vector cipher text and returns the result cipher text to the client. The third stage is the result process stage, it mainly decrypts the returned results, processes the decrypted results, and then sorts the processing results according to the similarity. The results of analysis show that the scheme has greatly improved in the safety, precision and retrieval efficiency.