Font Size: a A A

Research On Robust Retrieval Algorithms For Encrypted Medical Volume Data Based On 3D Transform Domain

Posted on:2019-05-21Degree:MasterType:Thesis
Country:ChinaCandidate:S S WangFull Text:PDF
GTID:2428330545993635Subject:Information and Communication Engineering
Abstract/Summary:PDF Full Text Request
Medical image data is an important auxiliary tool for doctors' diagnosis and medical research.Along with the continuous improvement of medical level,medical imaging technology has also developed rapidly,the number of digital image data in hospitals has increased dramatically.Problems such as the management and storage of medical image data also followed.Thanks to the development of cloud computing and big data,hospitals can store medical image data on cloud platforms to save costs.However,as an open third-party service provider,the cloud storage platform is not completely trusted,which may lead to information leakage and raise concerns about the safety of medical information.Therefore,how to simultaneously take into consideration the dual requirements of safe retrieval of patient privacy data and efficient and rapid retrieval,which has become an urgent problem to be solved in current encrypted medical image retrieval domain.At present,most of the multi-modality image data generated by medical imaging equipment are three-dimensional medical images.While,most of the current research focuse on two-dimensional images.There are few studies on three-dimensional medical images and which in ciphertext domain.In response,the paper has mainly research on three parts as follows:1)Studied a robust retrieval algorithm for encrypted medical volume data based on wavelet transform and chaotic map.It includes three processes:medical volume data encryption,feature extraction,and retrieval of encrypted medical volume data.The.algorithm combines the multi-resolution feature of wavelet transform and the good performance of Logistic map in the image encryption stage to ensure the security of the algorithm.In the process of image retrieval,the cloud server extract feature without knowing the content of the image,and then compares it with the feature vector in the cloud database,and returns an encrypted image with a high degree of similarity to the query image,then the retrieval process is achieved.Experimental results show that the algorithm has good retrieval accuracy and robustness.2)Studied an encrypted medical data robust retrieval algorithm based on DFT transform and perceptual hash.In this algorithm,the robustness and abstractness of image perceptual hash provide a foundation for the efficient and accurate retrieval of medical images.The encryption of medical volume data combines DFT transform and Logistic mapping to improve the security of the algorithm.The cloud server extracts the perceptual hash value of the encrypted medical volume data,and then calculates correlation coefficients one by one with the perceptual hash value of the encrypted image database stored in the cloud server,then complete the retrieval process.Experimental results show that the algorithm has good robustness and retrieval efficiency while satisfying privacy protection requirements.3)Researched a robust algorithm for encrypted medical volume data retrieval based on wavelet transform and perceptual hash.Considering that high-dimensional chaos has higher complexity and higher iteration speed than one-dimensional chaos,the algorithm uses Henon Map to protect medical information security.The algorithm combines good robustness of the perceptual hash function and extracts perceptual hash value of the encrypted volume data.When the encrypted volume data is deformed,it can still achieve fast and efficient retrieval.The algorithm comparison and experimental data show that the algorithm has good robustness against a certain intensity attacks under the premise of realizing retrieval.At the same time,the retrieval time is short and can meet the needs of daily use.
Keywords/Search Tags:encryption medical volume data, chaotic map, robust retrieval, feature vectors, perceptual hashing
PDF Full Text Request
Related items