Wildcard searchable encryption allows the server to efficiently perform wildcard-based keyword searches over encrypted data while maintaining data privacy. A promising solution to achieve wildcard SSE is to extract the characteristics of the queried keyword and check the existence based on a membership test structure. However, existing schemes have false positives of character order, that is, the server cannot identify the order between the first and the last wildcard character. Besides, the schemes also suffer from characteristic matching pattern leakage due to the one-by-one membership testing. In this paper, we present the first efficient wildcard SSE scheme to eliminate the false positives of character order and characteristic matching pattern leakage. To this end, we design a novel characteristic extraction technique that enables the client to exact the characteristics of the queried keyword maintaining the order between the first and the last wildcard character. Then, we utilize the primitive of Symmetric Subset Predicate Encryption, which supports checking if one set is a subset of another in one shot to reduce the characteristic matching pattern leakage. Finally, by performing a formal security analysis and implementing the scheme on a real-world database, we demonstrate that the desired security properties are achieved with high performance.

Symmetric Searchable Encryption (SSE) enables querying over encrypted data stored on servers, protecting the privacy of both client queries and the data. However, the response results of the query can leak certain statistical information, allowing an active attacker to inject files and exploit these leakages to recover the queried keyword. Recently, Zhang et al. (USENIX Security 2023) proposed a volumetric injection attack, called BVMA, based on response volume patterns. This attack requires larger file injections and has a moderate recovery rate. In this work, we propose a powerful volumetric injection attack,$\mathsf {Violin}$, that utilizes response length and size patterns, achieving a higher recovery rate and smaller injection size. Specifically,$\mathsf {Violin}$requires$\mathcal {O}(\log n)$injection files with a total size of$\mathcal {O}(n\log n)$, where$n$is the number of keywords. Furthermore, we propose a file injection method with optimal injection amount (i.e., injected files number and size) based on Pascal’s triangle rule under threshold countermeasure. We implement$\mathsf {Violin}$and compare it with the existing injection attacks. Experimental results indicate that$\mathsf {Violin}$achieves a recovery rate approaching 100%, surpassing the state-of-the-art BVMA while requiring smaller file sizes. Under threshold countermeasure,$\mathsf {Violin}$achieves a saving of 22% in injection file size and a remarkable 95% decrease in injected file number compared to BVMA.

Multi-sequence MRI image fusion technology aids radiologists in quickly and accurately assessing ischemic lesions and their surrounding areas by combining DWI and FLAIR images to generate information-rich fusion images. Despite the rapid development of medical image fusion techniques, existing methods are predominantly focused on technical-level model optimization and fail to effectively integrate medical domain knowledge. This limitation reduces their clinical applicability and model interpretability. Inspired by radiologists' diagnostic pattern, which involves focusing on and enlarging lesion areas, we propose a medical domain knowledge-inspired area amplification network for multi-sequence MRI image fusion in ischemic stroke, named MDKFusion. Specifically, we design the Lesion Area Amplification (LAA) module, which uses bicubic interpolation for adaptive amplification and incorporates crosslevel and neighboring-level feature mapping with high-level feature co-guidance. This design emulates radiologists' practice of zooming in to examine lesions, thereby enhancing interpretability. Additionally, we employ the Feature Guidance Module (FGM) to achieve progressive guidance and feature integration. We further introduce the ℒSCDloss function to minimize pixel discrepancies between source and fused images, improving fusion quality. Compared to various mainstream fusion methods, MDKFusion achieves state-of-the-art (SOTA) performance across eight objective evaluation metrics. To confirm its practical value in clinical diagnosis, we invited five radiologists to perform a subjective evaluation of the fused images. Our code will be available at https://github.com/MinLila/MDKFusion.

Secure join queries over encrypted databases, the most expressive class of SQL queries, have attracted extensive attention recently. The state-of-the-art JXT (Jutla et al. ASIACRYPT 2022) enables join queries on encrypted relational databases without pre-computing all possible joins. However, JXT can merely support join queries over two tables (in encrypted databases) with some high-entropy join attributes. In this paper, we propose an equi-join query protocol over two tables dubbed JXT+, that allows the join attributes with arbitrary names instead of JXT requiring the identical name for join attributes. JXT+ reduces the query complexity from O(\ell_1 \cdot \ell_2) to O(\ell_1) as compared to JXT, where \ell_1 and \ell_2 denote the numbers of matching records in two tables respectively. Furthermore, we present JXT++, the first equi-join queries across three or more tables over encrypted databases without pre-computation. Specifically, JXT++ supports joins of arbitrary attributes, i.e., all attributes (even low-entropy) can be candidates for join, while JXT requires high-entropy join attributes. In addition, JXT++ can alleviate sub-query leakage on three or more tables, which hides the leakage from the matching records of two-table join. Finally, we implement and compare our proposed schemes with the state-of-the-art JXT. The experimental results demonstrate that both of our schemes are superior to JXT in search and storage costs. In particular, JXT+ (resp., JXT++) brings a saving of 49% (resp., 68%) in server storage cost and achieves a speedup of 51.7× (resp., 54.3×) in search latency.

Encrypted conjunctive search enables server to perform efficient conjunctive query over encrypted data while guaranteeing data and query privacy. The well-known Oblivious Cross-Tags (OXT) protocol (by Cash et al. in CRYPTO 2013) is the first to realize efficient conjunctive search with some well-defined leakage, such as the keyword pair result pattern (KPRP) leakage and the cross-query intersection result pattern (IP) leakage. To mitigate the potential threats brought by the leakage, much effort has been made to reduce the information leaked by OXT. However, it is still open to achieve encrypted conjunctive search without revealing both KPRP and IP, while preserving high-efficiency.

Order-revealing encryption (ORE) enables the untrusted server to perform greater-than-comparison over ciphertext without compromising data privacy, which allows anyone to evaluate the lexicographic ordering of two arbitrary ciphertexts with a public comparison algorithm. However, most ORE constructions merely support ciphertext comparison for single-user. Recently, a variant of ORE named delegatable ORE has been introduced, which achieves cross-user ciphertext comparison by employing token mutual authorization technique at the cost of weak security, i.e., reveals the most significant differing bit of underlying plaintexts. To tackle this problem, we first present a deterministic property-preserving hash called DPPH with short-size hash value, and then propose a novel multi-client ORE scheme (m-ORE) from DPPH that supports ciphertext comparison among multiple users while hiding the most significant differing bits. Furthermore, we present an enhanced construction dubbed m-H-ORE by introducing a two-phase comparison method, which can achieve supper-efficient comparison in some cases, i.e., two ciphertexts with different bit-length. Finally, we provide formal security proofs of the proposed schemes and run extensive experiments to evaluate their performance on real-world and synthetic datasets. The results demonstrate that both of the proposed schemes can achieve a speedup of 47× and 138× in comparison cost to that of parameter-hiding ORE, respectively.

With the widespread development of intelligent systems, a considerable number of mobile devices are connected together, and continuously generate huge amounts of data. Although cloud storage provides perfect solution for effectively storing these massive data, how to ensure the integrity of the outsourced data becomes challenging. For this reason, the primitive of verifiable data streaming (VDS) protocol was introduced, and enables a data owner to continuously outsource streaming data to an untrusted cloud server, while capturing the integrity of the outsourced data. That is, when a data user retrieves some data item via its index from the server, he/she can publicly verify its integrity with the proof generated and returned by the server. Supporting data update is one of the major features of VDS, and allows the data owner to replace an old data item with a new one. Although many VDS protocols have been proposed to enhance the functionality and efficiency of the original VDS protocol, they all ignore the issue of preserving those updated data items. In fact, in various application scenarios of VDS, preserving and storing previously updated data items is actually necessary. For example, in the setting of DNA sequencing, there might be multiple versions of DNA fragments at the same location due to the genetic mutation. Obviously, for more precise treatment, all these DNA fragments need to be preserved. To this end, in this paper, we propose a VDS protocol that features of enabling the query of the update history of each data item. Specifically, we first put forward a new chameleon authentication tree with update history (UCAT), which consists of two CATs (the basic tree and the update history tree). In more detail, the basic tree is used to store the data item appended to the corresponding location for the first time, and the update history tree is utilized to preserve each updated version of the corresponding data item. Furthermore, based on UCAT, we propose a VDS protocol supporting update history queries, which allows a data user to retrieve any version of the data item. The theoretical analysis and performance evaluation indicate that our protocol outperforms previous ones in the field of functionality, and its computation/communication costs are acceptable. We also prove its security in the standard model.

The cloud-assisted intelligent systems have attracted extensive attention due to their powerful data analysis and computation capabilities. However, how to handle encrypted data remains a challenging problem in intelligent systems. A promising solution is searchable symmetric encryption (SSE), which enables a client to privately outsource their data to the cloud while preserving keyword search functionality. In practice, dynamic SSE is more practical and supports efficient data addition and deletion. Unfortunately, data update will leak some additional information which can be exploited to break data privacy. To address this issue, forward and backward secure SSE schemes are proposed to reduce the leakage of data update. That is, forward security guarantees that the newly updated documents cannot reveal the previously searched keywords, while backward security guarantees that the server cannot recover the deleted documents. However, the existing forward and backward secure SSE schemes mainly consider curious-but-honest server. How to verify the soundness and completeness of search results is still a challenge. In this paper, we propose a noninteractive verifiable dynamic SSE scheme with forward and backward security from two universal accumulators. Specifically, the server in our scheme only needs one roundtrip to return the nondeleted search results to the client, which saves the communication overhead dramatically. Besides, our scheme can achieve public verification that anyone can verify the search results but not only the client who has the private key. Finally, we give a formal security analysis and compare the proposed scheme with other related work, the results show that our scheme can achieve the desired security properties with practical efficiency.

Searchable Encryption (SE) enables a data owner to outsource encrypted data to an untrusted server while preserving the keyword search functionality. Typically, the server learns whether or not a query has been performed more than once, which is usually called the search pattern. However, such kind of information leakage might be leveraged to break query privacy. To further reduce such type of leakage and provide strong privacy guarantee, Wanget al.proposed a novel SE scheme based on the Paillier encryption scheme in INFOCOM’15. Unfortunately, their scheme cannot perform keyword search successfully, because the additive homomorphic property is not sufficient for their construction. In this article, we first show that why their scheme fails to return the correct search result, and then propose a new SE scheme by adopting a special additive homomorphic encryption scheme to achieve the multiplicative homomorphic property efficiently. Furthermore, we enhance the security on the user side. Specifically, we use random polynomials with an appropriate degree to guarantee that the user cannot learn anything other than the desired search result. Finally, we present a formal security analysis and implement our scheme on a real-world database, which demonstrates that our construction can achieve the desired security properties with good performance.