藉助同態加密方案的隱私保護病症檢索系統

Privacy Preserving Symptoms Retrieval System with the Aid of Homomorphic Encryption Schemes

王議偉(Yi-Wei Wang)

指導教授 ： 吳家麟

國立臺灣大學/電機資訊學院/資訊工程學研究所/碩士(2021年)

https://doi.org/10.6342/NTU202101681

摘要

本論文旨在提出一個系統，其能夠在確保各醫院病歷資料庫不被洩露的情況下，讓醫院彼此間可分享就診病患之病症資訊，且調閱資料病患的隱私將會被保護。在此系統中，我們假設醫院的安全等級是semi-honest (curious but honest），每間醫院需要聘任一位可信任的病歷室管理人 (Administrator），負責管理由其他醫院所取得的病患就診資訊。
我們提出的系統不僅能夠保護病患隱私並讓醫師跨院取得病患的病症資料，還能預防中間人攻擊 (Man in the Middle Attack） 以及內部攻擊 (Insider Attack） 等威脅。
我們的系統在單執行緒 (Single thread） 執行之下，向存有100萬筆病歷的醫院調閱 10000 筆病症僅需不到30秒，向存有1600萬筆病歷的醫院調閱 10000 筆病症僅需不到 6 分鐘，若以多執行緒 (Multi-thread） 執行，預期可在1秒左右調閱完畢。
我們的系統不只可以檢索病症資料，事實上病症之細項資料只要經過適當的編碼後也能夠調閱，例如: 糖尿病指數、高血壓指數、經非對稱數字系統 (Asymmetric Numeral Systems, ANS） 壓縮過的X光片、醫囑等，若要在醫院以外的情境使用應該也可以適用，例如醫療保險所牽涉到的病患銀行帳戶系統。

關鍵字

隱私保護集合交集 ； 同態加密 ； 複雜度降級 ； 數位簽章 ； 系統安全性分析

並列摘要

This work proposes a system allowing hospitals to share private information of patients while ensuring that each hospital's medical record database is not leaked, and the privacy of patients who access the data will be protected. We assume that the security level of the hospital is semi-honest, and each hospital needs to hire a trusted medical records department administrator to manage the private information of patients obtained from other hospitals.
Our proposed system not only protects patient privacy and allows physicians to obtain patient information across hospitals but also prevents threats such as man-in-the-middle attacks and insider attacks.
Our current system is executed on a single thread. It takes less than 30 seconds to retrieve the symptom information of 10,000 patients from a hospital with 1 million medical records and less than 6 minutes to retrieve the symptom information of 10,000 patients from a hospital with 16 million medical records. If the system were executed under multiple threads, the symptom information retrieval is expected to be completed within 1 second.
Our system can retrieve the symptom information and access the detailed information of the disease as long as it is adequately recorded in the form of numbers. For example, the diabetes index, hypertension index, medical X-ray film compressed by Asymmetric Numeral Systems (ANS), and doctor’s orders compressed by ANS can all be represented by fix-point numbers. Such a system could also be applied even if used in situations outside the hospital, such as the patient’s bank account system involved in medical insurance.

並列關鍵字

Private Set Intersection ； Homomorphic Encryption ； Complexity Reduction ； Digital Signature ； System Security Analysis

參考文獻

[1] R. Rivest, L. Adleman, and M. Dertouzos. (1978) "On data banks and privacy homomorphisms". Foundations of secure computation, 4(11):169−180.

[2] S. Goldwasser, S. Micali (1982). "Probabilistic encryption and how to play mental poker keeping secret all partial information". Proc. 14th Symposium on Theory of Computing: 365–377.

[3] D. Boneh, E. Goh, and K. Nissim (2005). "Evaluating 2-DNF Formulas on Ciphertexts". In Theory of Cryptography Conference.

[4] Craig Gentry (2009). "Fully homomorphic encryption using ideal lattices". In STOC, volume 9, pages 169−178.

[5] Zvika Brakerski, Craig Gentry, and Vinod Vaikuntanathan (2012). "(leveled) fully homomorphic encryption without bootstrapping". In Proceedings of the 3rd Innovations in Theoretical Computer Science Conference, pages 309–325. ACM.