Proceedings of the 2nd International Conference on Intelligent Data Communication Technologies and Internet of Things (IDCIoT-2024)

IEEE Xplore Part Number: CFP24CV1-ART; ISBN: 979-8-3503-2753-3

Homomorphic Encryption for Secure Data Analysis:

A Hybrid Approach using PKCS1_OAEP Padding
2024 2nd International Conference on Intelligent Data Communication Technologies and Internet of Things (IDCIoT) | 979-8-3503-2753-3/24/$31.00 ©2024 IEEE | DOI: 10.1109/IDCIOT59759.2024.10468008

Shanmukha Aditya G Kruthika B Shinu M .Rajagopal

Department of Computer Science and Department of Computer Science and Department of Computer Science and
Engineering, Amrita School of Engineering, Amrita School of Engineering, Amrita School of
Computing, Amrita Vishwa Computing, Amrita Vishwa Computing, Amrita Vishwa
Vidyapeetham, Bengaluru, India Vidyapeetham, Bengaluru, India Vidyapeetham, Bengaluru, India
[email protected] [email protected] [email protected]

C. R. Kavitha
Department of Computer Science and
Engineering, Amrita School of
Computing, Amrita Vishwa
Vidyapeetham, Bengaluru, India
[email protected]

Abstract—In today's data-driven world, striking a balance encryption underscore its significance in advancing a
between data utility and privacy is paramount. Homomorphic responsible and secure data analytics landscape. One notable
encryption, a revolutionary cryptographic technique, facilitates approach involves the application of Partial Ho mo morphic
computations on encrypted data without decryption. This study Encryption (PHE), featuring a dual-layered encryption
addresses the growing demand for secu re data analysis, paradigm and user-generated keys to bolster the privacy and
emphasizing the seamless integration of homomorphic integrity of medical data [1].
encryption in connecting robust data analysis with essential
privacy safeguards. Serving as a vital component in this context, In the rapidly evolving landscape of secure cloud
this cutting-edge technique proves pivotal in fostering a secure, computing, various cryptographic frameworks have been
privacy-centric approach to data analytics. The outcomes proposed to safeguard sensitive data and uphold privacy
underscore its significance, representing a substantial stride standards. Extending the scope, a secure cloud-based e-
towards the evolution of a conscientious and responsible data- voting system leverages Fully Ho mo morphic Encryption
driven ecosystem. The achieved results signal progress in with Chinese Remainder Theorem (FHECRT) to encrypt and
harmonizing data utility and privacy, enhancing overall security process votes, ensuring privacy and minimizing tampering
and ethical data practices without compromising analytical risks, alongside a comparative analysis with other
capabilities. homomorphic encryption algorithms [2]. Investigating the
Keywords— AES Encryption, Cloud Security, Homomorphic intersection of security and the Internet of Things (IoT), a
Encryption, Privacy Preservation, RSA Encryption, Secure Data study explores the feasibility of implementing homo morphic
Analytics encryption on IoT devices, scrutinizing resource
consumption, and offering optimization guidelines using
SEA L and HElib libraries on the Raspberry Pi 4 platform
I. INT RODUCT ION [3].
Addressing the challenges of security in IoT
In the era of pervasive data utilization, homo morphic environments, another research endeavor proposes a hybrid
encryption emerges as a transformative cryptographic data aggregation scheme utilizing symmetric encryption
paradigm, allowing computations on encrypted data without (Advanced Encryption Standard (AES), Triple Data
the need for decryption. This groundbreaking technique Encryption Standard (DES)) and ho mo morphic encryption,
addresses the pivotal challenge of balancing data utility and aiming to elevate security, accuracy, and reduce latency
privacy in this data-driven world. Ho mo morphic encryption while prioritizing privacy [4]. In the realm of hybrid
ensures that sensitive information remains confidential encryption models, a co mprehensive study introduces and
during computational processes, fostering a secure and implements four models that combine AES and Blowfish for
privacy-centric approach to data analysis. Its importance lies input encryption, coupled with Rivest, Shamir, Adleman
in mitigating the inherent trade-off between data accessibility (RSA) and Elliptic Curve Cryptography (ECC) for key
and protection, providing a means to extract valuable insights exchange. The performance and randomness of these models
without compromising confidentiality. Applications of are meticulously evaluated using various metrics [5].
homomorphic encryption span various domains, including Additionally, a co mparative analysis investigates the
healthcare, finance, and secure cloud computing. In efficiency of symmetric (A ES, DES, Caesar Cipher, Stream
healthcare, for instance, it enables collaborative research on Cipher) and asymmetric (Diffie -Hellman, RSA)
encrypted patient data, preserving individual privacy. In cryptographic algorithms, with a particular focus on file and
finance, it allows secure computation on encrypted financial image encryption, providing valuable insights into
transactions. The versatile applications of homomorphic processing time and CPU usage [6]. This diverse spectrum of

979-8-3503-2753-3/24/$31.00 ©2024 IEEE 481

Authorized licensed use limited to: ULAKBIM UASL - Osmangazi Universitesi. Downloaded on December 06,2024 at 16:11:34 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the 2nd International Conference on Intelligent Data Communication Technologies and Internet of Things (IDCIoT-2024)
IEEE Xplore Part Number: CFP24CV1-ART; ISBN: 979-8-3503-2753-3

research underscores the multifaceted nature of Arithmetic Circuit Ho mo morphic Encryption (ACHE),
cryptographic solutions in securing cloud-based systems and comparing it with Elliptic Curve Diffie Hellman (ECDH)
IoT environments. and analyzing factors such as latency, key generation,
encryption, and CPU usage. Evaluating key and ciphertext
The proposed model stands out by ensuring a secure data sizes, [14] assesses Paillier, ElGamal, ASHE, and Sy mmetric
import process from service seekers. Upon importing, the
for secure cloud computing, with a specific emphasis on low-
data is promptly encrypted within the data lake, a pivotal step memory devices. In a novel application, [15] proposes using
in safeguarding confidentiality, integrity, and accountability.
Homo morphic Encryption with GPS to enhance location
The stringent measures implemented cut off download
privacy for VIPs in real-t ime, encrypting and sharing journey
permissions for all parties—1st, 2nd, and 3rd, assuring that details confidentially. Furthermore, [16] introduces DPP, a
once uploaded, raw data remains inaccessible.
homomorphic encryption-based scheme for secure cloud
The major contributions of this study are: computing, ensuring data privacy in asynchronous
transactions.
 This model is poised to revolutionize how
Multinational Corporations (MNCs) handle Identifying vulnerabilities, [17] demonstrates known
proprietary information, guaranteeing its safety plaintext attacks on the Omar and Abed homo morphic
and accessibility from any location, at any time. encryption scheme. In the realm of b lockchain privacy, [18]
introduces a strategy utilizing homo morphic encryption for
 Notably, the proposed approach integrates data secure data sharing, addressing privacy concerns, and
visualization and analysis directly within the enhancing accuracy and efficiency in data transmission. For
data lake, eliminating the need for data plaintext encryption in homo morphic encryption, suggests
movement, copying, or sharing among different an improved paradigm, outsourcing computation to a
entities. delegated proxy to alleviate the co mputational burden on
 The strategic design of the study ensures that lightweight terminals in stream data applications.
sensitive information remains within the secure Additionally, [19] proposes a privacy-preserving logistic
enclave of the data lake, min imizing potential regression scheme based on leveled fully ho mo morphic
vulnerabilities and fort ifying overall data encryption (CKKS) and trusted hardware, enabling efficient
security. model training on encrypted data while preserving data
privacy.
Section II offers an essential overview of homo morphic
encryption, including a review of relevant prior research. Presenting an innovative privacy-preserving logistic
Section III presents an in-depth analysis of the dataset, regression scheme, [20] leverages trusted hardware and
including its orig in and detailed description. Section IV is Leveled FHE for encrypted computations, offloading
dedicated to the research methodology employed. Section V bootstrapping and non-linear activation functions to trusted
elaborates on the findings and provides a discussion on the hardware for enhanced efficiency and accuracy.
implications of the results. Lastly, Section VI concludes the Acknowledging the challenges of current homomorphic
paper, offering a summary and highlighting potential encryption, [21] proposes harnessing hardware like FPGAs
avenues for future research in this field. to turbocharge computations, paving the way for faster,
privacy-preserving cloud computing. [22] contributes to the
literature by reviewing existing homo morphic encryption
protocols for secure cloud data processing, proposing a new
II. RELAT ED WORK protocol with security analysis and experiments, and
emphasizing the need for mo re practical schemes for real-
In this retrospective analysis, classical homo morphic world cloud applications.
encryption algorithms, including RSA, El-Gamal, and Enhancing the first identity-based fully ho mo morphic
Paillier, are scrutinized, emphasizing their properties and encryption (IBFHE) scheme by leveraging improved noise
evolution towards achieving fully homo morphic encryption analysis methods and novel trapdoors, and introduces an
[8]. In a unique approach, [9] co mbines ITU-T X.690 Basic efficient mu lti-identity fully homo morphic encryption
Encoding Rule with Arithmetic Circuit Ho mo morphic (MIFHE) scheme using a masking approach, enabling
Encryption for secure computation of integer expressions, homomorphic evaluation of ciphertexts under different
delving into addition, subtraction, and multiplication across identities [24].
nodes, and providing insights into protocol design and
latency analysis. Addressing the panorama of fully The study critiques a symmetric Fu lly Ho mo morphic
homomorphic encryption schemes, [11] conducts a thorough Encryption (FHE) scheme based on Euler's theorem,
analysis and comparison, encompassing lattice-based, exposing a vulnerability allowing secret key recovery,
integer-based, ring learning with error, and elliptic curve proposes an improved scheme based on the approximate
cryptography-based schemes, discussing their properties, greatest common divisor (A GCD) problem, and assesses
security, efficiency, and potential improvements. various attacks, including those introduced in the paper [25].
Furthermore, [12] introduces a modification to Keyed
A distinctive aspect of this research study lies in its
Homo morphic-Public Key Encryption (KH-PKE) to enhance
commit ment to data security and privacy. Unlike existing
security in cloud-based machine learning scenarios, ensuring
studies, the methodology ensures that once uploaded, the raw
that only authorized users, specifically the client, can encrypt
data remains inaccessible even to the first party. This
and decrypt data.
emphasis on robust data protection constitutes a novel
Shifting focus to post-quantum security, [13] imp lements contribution, fortifying the integrity of this research and
Super singular Isogeny Diffie Hellman Key Exchange in setting a precedent for a more secure approach to data

979-8-3503-2753-3/24/$31.00 ©2024 IEEE 482

Authorized licensed use limited to: ULAKBIM UASL - Osmangazi Universitesi. Downloaded on December 06,2024 at 16:11:34 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the 2nd International Conference on Intelligent Data Communication Technologies and Internet of Things (IDCIoT-2024)
IEEE Xplore Part Number: CFP24CV1-ART; ISBN: 979-8-3503-2753-3

handling in the realm of machine learning and data analysis.

This unique feature underscores the dedication to advancing
not only the analytical aspects but also the overarching IV. M ET HODOLOGY
principles of data confidentiality.
Cryptographic techniques play a pivotal role in ensuring
the privacy and security of sensitive data, especially in the
III. DAT A SOURCE AND DESCRIPT ION context of machine learning datasets that often contain
personally identifiable information (PII). This section
delineates a detailed methodology employed in a research
The dataset utilized for this research project is the widely
endeavor focused on leveraging cryptographic algorithms,
recognized Adult Income dataset [23], extensively employed
specifically RSA and A ES, to enhance data privacy. The
in K-nearest neighbors (KNN) analyses, making it an ideal
mu ltifaceted methodology encompasses data encryption,
choice for investigating the application of homomorphic
decryption, and subsequent analysis, providing a holistic
encryption in secure data analysis. This dataset serves as a
approach to safeguarding sensitive information as shown in
fundamental foundation for exploring the intricacies of
Fig.2.
homomorphic encryption, ensuring that sensitive information
remains confidential during analysis. With 16 columns, the
dataset's central focus is the "Income" target field,
categorizing individuals into two classes: those earning less
than or equal to $50,000 (<=50K) and those earning more
than $50,000 (>50K), as depicted in Fig. 1.
The dataset boasts 14 attributes, providing a
comprehensive set of features that delve into the
demographics and personal characteristics of individuals.
Crucial factors such as education level, age, gender, and
occupation are included, offering a rich source of
information for predictive modeling. This diversity of
Fig.2. Architectural diagram of the proposed Homomorphic Encryption
attributes contributes to the dataset's suitability for examining model for Secure Data Analysis
the effectiveness of homomorphic encryption in preserving
the privacy and security of individual-level information. A. Data Encryption
Each dataset entry encapsulates an individual's annual
income, offering a nuanced perspective influenced by
various factors. The initial phase of the methodology focuses on the
encryption of sensitive information within the dataset, which
The dataset's appropriateness for this project lies in its is critical for protecting data during storage or transmission
real-world relevance and the complexity of its features, and mitigating the risk of unauthorized access. The
making it an excellent candidate for evaluating the cryptographic approach employed is a hybrid co mbination of
practicality and efficacy of ho mo morphic encryption in a RSA and A ES algorith ms, leveraging the unique strengths of
secure data analysis context. The application of each to fortify the security of the dataset.
homomorphic encryption to such a dataset enables the
exp loration of innovative approaches to maintaining privacy The encryption process begins with the generation of an
while extracting meaningful insights, aligning with the RSA key pair, co mprising a public key for encryption and a
project's overarching goal of enhancing data security in the private key for decryption. The public key is stored for future
context of sensitive information analysis. cryptographic operations. Each data cell within the dataset is
then transformed: its content is converted to bytes and
padded to a mu ltiple of 16 bytes, as required by the AES
algorithm.
AES, a symmetric-key algorithm, is used to encrypt the
data. Randomly generated AES keys and initialization
vectors (IV) enhance security. The data encryption is
performed using the AES cipher with the derived key and
IV. Concurrently, the AES key and IV are encrypted using
the RSA public key. This results in a unified encrypted
message that combines the encrypted AES key, IV, and data,
converted into a hexadecimal string for efficient storage and
retrieval.
The role of RSA in this process is crucial. RSA,
particularly with PKCS1_ OAEP padding, provides a secure
and effective means of encrypting the AES keys and IV.
PKCS1_OA EP is a sophisticated padding scheme that
significantly enhances RSA security by incorporating
randomness and a structured padding mechanism. This
Fig.1. Income vs count graph corresponding to the dataset to show the scheme is vital in preventing chosen ciphertext attacks and
instances of the dataset. ensuring the confidentiality and integrity of the encrypted

979-8-3503-2753-3/24/$31.00 ©2024 IEEE 483

Authorized licensed use limited to: ULAKBIM UASL - Osmangazi Universitesi. Downloaded on December 06,2024 at 16:11:34 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the 2nd International Conference on Intelligent Data Communication Technologies and Internet of Things (IDCIoT-2024)
IEEE Xplore Part Number: CFP24CV1-ART; ISBN: 979-8-3503-2753-3

data. It adds an additional security layer by making the RSA RSA and AES algorithms in a hybrid approach effectively
encryption process more resilient against various transformed the dataset into a secure, encrypted format. This
cryptographic attacks and vulnerabilities typically associated method was crucial in maintaining the confidentiality of
with simpler padding schemes. sensitive information during storage and transmission. In the
decryption phase, the employment of PKCS1_ OAEP
B. Data Decryption padding with RSA decryption proved efficient and secure in
retrieving the orig inal data structure. Maintaining the
The decryption phase is essential for ensuring that integrity of decrypted data is vital for trustworthy and
authorized entities can securely access and utilize the reliable secure data analysis.
sensitive informat ion. It involves reversing the encryption
steps to yield a dataset that mirrors the original data's A distinctive feature of this model is its focus on secure
structure and content. data import and in-lake analysis, eliminating the need for
data movement and sharing among different entities and
The process begins with the extraction of the RSA enhancing the overall security framework. Th is approach is
private key, created during the RSA key pair setup and saved particularly beneficial for scenarios requiring high data
for later use. The decrypted AES key and IV are obtained by privacy, reducing the risk of data breaches and unauthorized
decrypting the corresponding components of the encrypted access. The model's practical implications are notably
message using the RSA private key. An AES cipher is then evident in the context of multinational corporations (MNCs).
constructed using the decrypted AES key and IV. It offers a novel solution for handling proprietary
information, ensuring safety, accessibility, and secure
The AES decryption step decrypts the encrypted data
analysis fro m any location. This global applicability
using the AES cipher, followed by the removal of any
highlights the model's versatility and potential to
padding introduced during encryption. This ensures the
revolutionize data handling practices across various sectors.
accurate retrieval of the original data and maintains the
integrity of the data during retrieval. To conclude with, the results from this Ho mo morphic
Encryption model signify a substantial advancement in
In this study, PKCS1_OAEP padding is also used in RSA
harmonizing data utility with privacy. It reinforces the
decryption, enhancing the security of the decrypted data by
development of a conscientious and responsible data-driven
preventing various cryptographic attacks. This randomized
ecosystem where data security and privacy are integral to
padding mechanism further strengthens the RSA algorithm,
analytical capabilities. This balance is crucial in the current
making it more resistant to potential vulnerabilit ies and
era of extensive data utilization, underscoring the importance
ensuring a secure and reliable decryption process.
and impact of this study in the realm of secure data analysis.

C. Handling the Decrypted data for Secure Analysis

VI. CONCLUSION AND FUT URE WORK

Upon successful decryption, the dataset is loaded into a

pandas Data Frame, allowing for seamless analys is and In conclusion, this research establishes a robust
exp loration of its contents. The integration of cryptographic framework for leveraging homomorphic encryption to
techniques enhances the data's security while preserving its enhance the privacy and security of sensitive data during
integrity throughout this process. The decrypted data, now in analysis. By implementing a hybrid encryption approach
a structured pandas Data Frame, can be utilized for various involving RSA and AES, the proposed architecture ensure
data analysis tasks. The generalized nature of this phase that data remains confidential throughout the encryption and
allows for the application of standard data analysis decryption processes. The model's uniqueness lies in its
techniques, machine learning models, or any other commit ment to data security, exemp lified by the restriction
exp loratory methods tailored to the specific research of raw data access even for the first party. The integration of
objectives. cryptographic techniques, such as Partial Ho mo morphic
Encryption (PHE), presents a pioneering solution fo r
Moreover, it is crucial to h ighlight that the decrypted securing medical data. Furthermore, this approach addresses
data, being transient and not stored in any persistent file, the challenges of secure cloud computing and IoT
eliminates the risk of raw or plain data lingering in the environments, emphasizing the versatile applications of
system. Th is design choice significantly mitigates potential homomorphic encryption.
security issues associated with data storage. Once th e
analysis is completed, the pandas Data Frame is promptly For future work, the study can be expanded by aiming to
deleted, ensuring that no remnants of the decrypted data exp lore advanced homomorphic encryption schemes,
persist within the system. This meticulous approach further investigate the scalability of the proposed model, and delve
fortifies the security posture, leaving no room for into real-world deploy ment challenges. Additionally,
unauthorized access to sensitive information. continuous refinement of encryption techniques and the
exp loration of emerging cryptographic protocols will be
crucial for sustaining the evolution of secure data analysis
methodologies.
V. RESULT S AND DISCUSSIONS

The imp lementation of the Homomorphic Encryption

model for Secure Data Analysis has demonstrated significant
advancements in the field of data security. Integrating both

979-8-3503-2753-3/24/$31.00 ©2024 IEEE 484

Authorized licensed use limited to: ULAKBIM UASL - Osmangazi Universitesi. Downloaded on December 06,2024 at 16:11:34 UTC from IEEE Xplore. Restrictions apply.
 Proceedings of the 2nd International Conference on Intelligent Data Communication Technologies and Internet of Things (IDCIoT-2024)
IEEE Xplore Part Number: CFP24CV1-ART; ISBN: 979-8-3503-2753-3

REFERENCES [13] W. X. K. Joshua, X. W. T. Justin and C. N. Yap, "Arithmetic Circuit

Homomorphic Encryption Key Pairing Comparisons and Analysis
between Elliptic Curve Diffie Hellman and Supersingular Isogeny
[1] G S. Murthy and C. R. Kavitha, "Preserving Data Privacy in Cloud Diffie Hellman," 2021 2nd Asia Conference on Computers and
using Homomorphic Encryption," 2019 3rd International conference Communications (ACCC), Singapore, 2021, pp. 138-142.
on Electronics, Communication and Aerospace Technology (ICECA), [14] J. Ryu, K. Kim and D. Won, "A Study on Partially Homomorphic
Coimbatore, India, 2019, pp. 1131-1135. Encryption," 2023 17th International Conference on Ubiquitous
[2] V. S. Anjima and N. N. Hari, "Secure Cloud e-Voting System Using Information Management and Communication (IMCOM), Seoul,
Fully Homomorphic Elliptical Curve Cryptography," 2019 Korea, Republic of, 2023, pp. 1-4.
International Conference on Intelligent Computing and Control [15] S. Gupta and G. Arora, "Use of Homomorphic Encryption with GPS
Systems (ICCS), Madurai, India, 2019, pp. 858-864. in Location Privacy," 2019 4th Internat ional Conference on
[3] H. M. Reddy, S. P C and S. Sankaran, "On the Feasibility of Information Systems and Computer Networks (ISCON), Mathura,
Homomorphic Encryption for Internet of Things," 2022 IEEE 8th India, 2019, pp. 42-45.
World Forum on Internet of Things (WF-IoT), Yokohama, Japan, [16] J. Wang, F. Wu, T. Zhang and X. Wu, "DPP: Data Privacy-Preserving
2022, pp. 1-6. for Cloud Computing based on Homomorphic Encryption," 2022
[4] G. K, M. S and H. K, "Hybrid approach for Data Aggregation in International Conference on Cyber-Enabled Distributed Computing
cloud," 2021 6th International Conference on Communication and and Knowledge Discovery (CyberC), Suzhou, China, 2022, pp. 29-32.
Electronics Systems (ICCES), Coimbatre, India, 2021, pp. 1-7. [17] S. Choi and H. T. Lee, "Known Plaintext Attacks on the Omar and
[5] A. P. J, K. Jain and R. Aragona, "Performance Comparison of Hy brid abed Homomorphic Encryption Scheme," 2022 13th International
Encryption Models," 2023 Second International Conference on Conference on Information and Communication Technology
Augmented Intelligence and Sustainable Systems (ICAISS), T richy, Convergence (ICT C), Jeju Island, Korea, Republic of, 2022, pp.
India, 2023, pp. 1196-1203. 1154-1157.
[6] N. K. S. Keerthan, S. P. Marri and M. Khanna, "Analysis of Key [18] L. Zhu, S. Song, S. Peng, W. Wang, S. Hu and W. Lan, "The
Based Cryptographic Algorithms and its Applications," 2023 IEEE Blockchain and Homomorphic Encryption Data Sharing Method in
3rd International Conference on Technology, Engineering, Privacy-Preserving Computing," 2022 IEEE/ACIS 7th International
Management for Societal impact using Marketing, Entrepreneurship Conference on Big Data, Cloud Computing, and Data Science (BCD),
and T alent (TEMSMET), Mysuru, India, 2023, pp. 1-4. Danang, Vietnam, 2022, pp. 84-87.
[7] C. Marcolla, V. Sucasas, M. Manzano, R. Bassoli, F. H. P. Fitzek and [19] X. Li, R. Li, B. Bai, Y. Zhao, G. Liu and R. Li, "Outsourcing the
N. Aaraj, "Survey on Fully Homomorphic Encryption, Theory, and Computation of Plaintext Encryption for Homomorphic Encryption,"
Applications," in Proceedings of the IEEE, vol. 110, no. 10, pp. 1572 - 2023 8th International Conference on Computer and Communication
1609, Oct. 2022. Systems (ICCCS), Guangzhou, China, 2023, pp. 408-413.
[8] S. Mittal and S. Singh, "Retrospective Study on Classical [20] C. Liu et al., "Efficient and Privacy-Preserving Logistic Regression
Homomorphic Encryption Algorithms," 2023 International Scheme based on Leveled Fully Homomorphic Encryption," IEEE
Conference on IoT, Communication and Automation Technology INFOCOM 2022 - IEEE Conference on Computer Communications
(ICICAT ), Gorakhpur, India, 2023, pp. 1-6. Workshops (INFOCOM WKSHPS), New York, NY, USA, 2022, pp.
[9] W. H. Soh Kenneth, C. J. Kwa and C. N. Yap, "Applying Basic 1-6.
Encoding Rule (ITU-T X.690) on Integer Expressions using [21] K. Aulakh and R. K. Ramachandran, "A Detailed Survey of Fully
Arithmetic Circuit Homomorphic Encryption," 2020 Asia Conference Homomorphic Encryption Standards to Preserve Privacy over Cloud
on Computers and Communications (ACCC), Singapore, Singapore, Communications," 2020 Indo – T aiwan 2nd International Conference
2020, pp. 91-95. on Computing, Analytics and Networks (Indo-Taiwan ICAN),
[10] Z. H. Mahmood and M. K. Ibrahem, "New Fully Homomorphic Rajpura, India, 2020, pp. 207-211.
Encryption Scheme Based on Multistage Partial Homomorphic [22] V. Swathi and M. P. Vani, "Privacy-Cheating Discouragement: A
Encryption Applied in Cloud Computing," 2018 1st Annual New Homomorphic Encryption Scheme for Cloud Data Security,"
International Conference on Information and Sciences (AiCIS), 2020 11th International Conference on Computing, Communication
Fallujah, Iraq, 2018, pp. 182-186. and Networking Technologies (ICCCNT), Kharagpur, India, 2020,
[11] P. Chaudhary, R. Gupta, A. Singh and P. Majumder, "Analysis and pp. 1-6.
Comparison of Various Fully Homomorphic Encryption Techniques," [23] https://www.kaggle.com/datasets/wenruliu/adult-income-dataset
2019 International Conference on Computing, Power and [24] T. Shen, F. Wang, K. Chen, K. Wang and B. Li, "Efficient Leveled
Communication Technologies (GUCON), New Delhi, India, 2019, (Multi) Identity-Based Fully Homomorphic Encryption Schemes," in
pp. 58-62. IEEE Access, vol. 7, pp. 79299-79310, 2019.
[12] T. Tsuruta, S. Araki, T . Miyazaki, S. Uehara and K. Kakizaki, "A [25] W. Xu, Y. Zhan, Z. Wang, B. Wang and Y. Ping, "Attack and
Study on a DDH-Based Keyed Homomorphic Encryption Suitable to Improvement on a Symmetric Fully Homomorphic Encryption
Machine Learning in the Cloud," 2022 IEEE International Conference Scheme," in IEEE Access, vol. 7, pp. 68373-68379, 2019.
on Consumer Electronics - T aiwan, T aipei, T aiwan, 2022, pp. 167-
168.

979-8-3503-2753-3/24/$31.00 ©2024 IEEE 485

Authorized licensed use limited to: ULAKBIM UASL - Osmangazi Universitesi. Downloaded on December 06,2024 at 16:11:34 UTC from IEEE Xplore. Restrictions apply.