GPU-Accelerated implementation of a genetically optimized image encryption algorithm
Open accessAbstract
This paper presents a GPU-accelerated implementation of an image encryption algorithm. The algorithm uses the concepts of a modified XOR cipher to encrypt and decrypt the images, with an encryption pad, generated using the shared secret key and some initialization vectors. It uses a genetically optimized pseudo-random generator that outputs a stream of random bytes of the specified length. The proposed algorithm is subjected to a number of theoretical, experimental, and mathematical analyses, to examine its performance and security against a number of possible attacks, using the following metrics - histogram analysis, correlation analysis, information entropy analysis, NPCR and UACI. The performance analysis carried out shows an average speedup-ratio of 3.489 for encryption, and 4.055 for decryption operation, between the serial and parallel implementations using GPU. The algorithm aims to provide better performance benchmarks, which can significantly improve the experience in the relevant use-cases, like real-time media applications.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
0 of 17 cited sources checked · 0 supported · 0 flagged
- Findingown data
The best coefficient of variation obtained by the genetic algorithm was 0.018369, which is lower than the 0.069726 reported for the algorithm of Bharadwaj and Sairabanu (2020).
“The COV value obtained for the best output of the genetic algorithm was observed to be 0.018369, which is much better than the COV value obtained for the algorithm proposed in Bharadwaj and Sairabanu (2020), which was observed to be 0.069726.”
Quote in the paperSubstantial claim0 of 1 sources checked- cites Bharadwaj and Sairabanu (2020) source
- Findingown data
The speedup ratio for decryption between the serial and GPU-parallel implementations is 4.055.
“The obtained values of the speedup ratio were 3.489 for encryption, and 4.055 for decryption, respectively.”
Quote in the paperSubstantial claim - Findingown data
The speedup ratio for encryption between the serial and GPU-parallel implementations is 3.489.
“The obtained values of the speedup ratio were 3.489 for encryption, and 4.055 for decryption, respectively.”
Quote in the paperSubstantial claim - Findingown data
The paper's comparison against earlier algorithms shows improved performance.
“A comparative analysis is also included with some of the previously proposed algorithms, and using quantifiable mathematical metrics, demonstrates the performance enhancements.”
Quote in the paperSubstantial claim - Findingown data
The encryption-decryption process caused no loss of visual information in the test images.
“As can be seen, there has been no loss of visual information in the entire encryption-decryption process.”
Quote in the paperSubstantial claim
- Findingown data
The correlation coefficients of the encrypted images are close to zero, indicating negligible correlation.
“It can be clearly seen that considerable levels of correlations are detected for the input images, whereas the same is in a close proximity to zero, for the encrypted images. This signifies a negligible degree of correlation.”
Quote in the paperSubstantial claim - Interpretationreasoning
The work builds on Bharadwaj and Sairabanu's (2020) encryption scheme by adding SIMD computing to improve efficiency benchmarks.
“This paper extends upon the ideas introduced by the encryption paradigm proposed in Bharadwaj and Sairabanu (2020) by incorporating the SIMD computing architecture into it, for achieving superior efficiency benchmarks.”
Quote in the paperSubstantial claim0 of 1 sources checked- cites Bharadwaj and Sairabanu (2020) source
- Interpretationnone given
The proposed algorithm uses a helical lattice structure with cryptographic convolutions to provide unidirectional properties.
“To achieve this, the algorithm proposes the architecture of a specially designed helical lattice structure, to incorporate cryptographic convolutions in the mathematics involved, in order to guarantee unidirectional attributes.”
Quote in the paperSubstantial claim - Interpretationreasoning
The convolution in the helical structure prevents an attacker who knows the output random-byte stream from computing the initial seed values or the shared secret key without brute force.
“This convolution operation is performed, in order to ensure that, even if the attacker somehow discovers the output stream of random bytes for any given process of encryption/decryption, he/she is still unable to compute the values of the initial seed values and the common secret key used for the computation, without resolving to the brute-force approach of trial and error.”
Quote in the paperNot a substantial claim- Rests on an assumption: The shared secret key remains secret from an attacker.
- Interpretationnone given
The convolution imposes an upper bound on the degree of parallelism achievable in the computations.
“However, this applies an upper-bound on the degree of parallelism that can be achieved in the involved computations, which can be addressed in future.”
Quote in the paperNot a substantial claim - Interpretationnone given
The algorithm prevents an attacker from recovering the seed values under a known-plaintext or chosen-plaintext attack unless the attacker knows the initial secret key.
“When the presented algorithm is subjected to a known-plaintext or chosen-plaintext attack, the attacker cannot decipher the seed values, without knowing the initial secret key (assumed secure for smooth operation).”
Quote in the paperSubstantial claim- The authors state this as an assumption: The two parties have mutually agreed on a common secret initial key, and that key remains secret throughout the algorithm's operation.
- Interpretationreasoning
The byte extraction operation never yields a zero random byte.
“It should be noted that this operation results in no random byte being 0”
Quote in the paperSubstantial claim - Interpretationreasoning
Each 64-bit entry in the Helix array yields seven random bytes.
“Using this operation, seven random bytes can be extracted from each 64-bit value present in the Helix data-structure.”
Quote in the paperNot a substantial claim - Interpretationreasoning
Future seed values are secure because they cannot be derived from decoded seeds without the secret key.
“By extension, the future seed values are also secure, as they can’t be obtained from the decoded seeds without the secret key.”
Quote in the paperSubstantial claim- Rests on an assumption: The shared secret key remains secret from an attacker.
- Interpretationnone given
If the initialization vector is unique for every run, then encrypting identical data with the same key produces different ciphertexts.
“It is also essential that the value of IV is unique for each and every run, which ensures that even if the same data is encrypted repeatedly, along with the same key, the obtained encryption outputs C1, C2, C3, etc. are still different.”
Quote in the paperSubstantial claim - Interpretationnone given
The PRNG generates bytes such that the same random bytes can be replicated at the receiver only when the correct inputs are supplied.
“These bytes are generated such that the same random-bytes can be replicated on the other end of the communication, only when the corresponding correct inputs are provided to it.”
Quote in the paperNot a substantial claim- Rests on an assumption: The shared secret key remains secret from an attacker.
- Premisecitation
GPU parallelism boosts the speed of the algorithm's SIMD-based components relative to older methods.
“The involvement of GPU in the algorithm allows it to make use of the parallel-computing architecture. This improves the performance of the paradigm’s parts that follow the SIMD structure, with respect to the traditional approaches that have been employed to solve the same problem. (Antwerpen 2011; Zafar et al. 2010)”
Quote in the paperSubstantial claim0 of 2 sources checked - Interpretationreasoning
Repeating the same plaintext does not help an attacker because the initialization vector is unique.
“Moreover, re-iterating the same plaintext would also not help, due to the uniqueness of the IV parameter.”
Quote in the paperSubstantial claim- The authors state this as an assumption: The initialization vector is unique for every run of the algorithm.
- May combine two claims. 12
- The quote only partly states this claim. 3
- Classified as probably not a substantial claim. 6
May combine two claims.
Findingown dataIn the authors' performance analysis, the parallel GPU implementation achieved an average encryption speedup ratio of 3.489 and an average decryption speedup ratio of 4.055 relative to the serial implementation.
“The performance analysis carried out shows an average speedup-ratio of 3.489 for encryption, and 4.055 for decryption operation, between the serial and parallel implementations using GPU.”
Quote in the paperSubstantial claimMay combine two claims.
Findingown dataThe histograms of the encrypted images show no contextual information, unlike the histograms of the input images.
“Upon examining these plots through Fig. 6, it was observed that even though some contextual information was visible from the histograms of the input images, there was no such feature visible in the same for the encrypted images.”
Quote in the paperSubstantial claimThe quote only partly states this claim.
Findingown dataThe proposed algorithm's average encryption and decryption speeds are far better than those of the compared schemes.
“It can be seen that the performance achieved using the proposed algorithm, is far better than the other schemes.”
Quote in the paperSubstantial claimMay combine two claims.
Findingown dataThe proposed algorithm's UACI values are at par with Ramesh and Jain (2015) and Bharadwaj and Sairabanu (2020), and better than Zhang et al. (2017).
“With respect to UACI values, it can be observed that the proposed algorithm works at-par with (Ramesh and Jain 2015) and (Bharadwaj and Sairabanu 2020), and has better results than (Zhang et al. 2017).”
Quote in the paperSubstantial claim0 of 3 sources checkedThe quote only partly states this claim.
Findingown dataThe entropy values of the encrypted images are at the expected level of 8.
“But, for the encrypted images, the entropy values are at par with the expected values.”
Quote in the paperSubstantial claimMay combine two claims.
Findingown dataThe proposed algorithm's entropy values are closer to 8 than those of Failed (2015) and Zhang et al. (2017), and comparable to those of Bharadwaj and Sairabanu (2020).
“the values are in a close agreement to those of Bharadwaj and Sairabanu (2020), but are quite closer to the desired value of 8, in comparison to the same values obtained using the algorithms from Failed (2015) and (Zhang et al. 2017), indicating better entropy levels for our algorithm.”
Quote in the paperSubstantial claim0 of 3 sources checkedMay combine two claims.
Findingown dataThe proposed algorithm's NPCR values are at par with Bharadwaj and Sairabanu (2020) and better than Ramesh and Jain (2015) and Zhang et al. (2017).
“It is quite clear that the proposed algorithm is performing at-par with (Bharadwaj and Sairabanu 2020), and better than (Ramesh and Jain 2015) and (Zhang et al. 2017), for NPCR values.”
Quote in the paperSubstantial claim0 of 3 sources checkedMay combine two claims.
Findingown dataThe proposed algorithm's correlation results are comparable to those of Bharadwaj and Sairabanu (2020) and better than those of Failed (2015) and Zhang et al. (2017).
“As can be seen, the proposed algorithm has comparable results to Bharadwaj and Sairabanu (2020), and outperforms (Failed 2015) and (Zhang et al. 2017).”
Quote in the paperSubstantial claim0 of 3 sources checkedThe quote only partly states this claim.
Interpretationnone givenApplying the decryption algorithm reverses the encryption modifications and returns an exact copy of the original image.
“The application of the above-mentioned algorithm, effectively nullifies the modifications introduced by the encryption process, thus providing the receiver with the exact copy of the initially encrypted image.”
Quote in the paperNot a substantial claimMay combine two claims.
InterpretationreasoningThe lower coefficient of variation produces a smoother histogram and therefore a higher security standard.
“This difference results in a smoother histogram for the proposed algorithm, and hence a higher security standard.”
Quote in the paperSubstantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
Interpretationnone givenThe algorithm's constants A, B, C, and D require tuning to reach optimal performance.
“It should be noted that the constants A, B, C, and D are hyper-parameters that need to be tuned, in order to obtain optimal results from the considered algorithm”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing), S4 (based).
Interpretationnone givenThe algorithm can operate with a key of arbitrary length, limited only by system-specific parameters such as memory, processing power, and required security.
“The presented algorithm is capable of functioning with a key of arbitrary length. It’s only constrained by the system-specific parameters, such as the available memory, processing power, required level of security, etc.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
Interpretationnone givenThe initialization vector and encryption vector are not secret and may be sent as plaintext over an insecure channel together with the encrypted image.
“Also note that the value of IV and EV is not secret, and can be transmitted as plain-text over an insecure channel, along with the encrypted image C.”
Quote in the paperNot a substantial claimProbably not a substantial claim: may not pass S4 (based).
Interpretationnone givenThe PRNG module provides a cryptographically secure set of random bytes efficiently.
“The crux of the entire encryption-decryption routine proposed in this paper is the PRNG module, which is responsible for providing a cryptographically-secure set of random-bytes in an efficient way.”
Quote in the paperNot a substantial claim- Rests on an assumption: The shared secret key remains secret from an attacker.
- Rests on an assumption: The chosen security metrics—histogram, correlation, information entropy, NPCR, and UACI—are sufficient to establish security against the considered attacks.
Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.
Interpretationnone givenThe algorithm does not explicitly provide protection against denial-of-service attacks; users must arrange such protection themselves.
“The algorithm does not explicitly provide a means to combat a Denial-of-Service (DoS) attack, for which suitable arrangements must be performed by the user, during the development of the system in consideration.”
Quote in the paperNot a substantial claimMay combine two claims.
InterpretationreasoningThe algorithm is secure against a man-in-the-middle attack provided the key-exchange mechanism authenticates the parties and the key space is at least the recommended size, because the computations are uni-directional and cannot be back-tracked to recover the seed values.
“For a Man-In-The-Middle (MITM) attack, as long as the key-exchange mechanism ensures that the authenticity of the parties involved in the communication is verified before transmission, and the key-space is of recommended size (or more), the algorithm is deemed secure.”
- Also stated in: § Analysis against various attacks · boosted
“It is so, because of the uni-directional nature of the involved computations, which make it mathematically impossible to replicate the seed values, by back-tracking the values in the algorithm.”
Quote in the paperSubstantial claim- Also stated in: § Analysis against various attacks · boosted
May combine two claims.
Interpretationown dataThe changes render the algorithm better suited to fast modern uses while keeping security levels intact.
“The improvements introduced in this paper make the proposed algorithm a lot more suitable for modern-day high-speed applications, without compromising on security parameters.”
Quote in the paperSubstantial claim- Rests on an assumption: The shared secret key remains secret from an attacker.
- Rests on an assumption: Sending the encryption vector EV with the ciphertext over an insecure channel does not let an attacker recover the plaintext or the secret key.
- Rests on an assumption: The chosen security metrics—histogram, correlation, information entropy, NPCR, and UACI—are sufficient to establish security against the considered attacks.
May combine two claims.
InterpretationreasoningBecause no random byte is zero, pattern-recognition algorithms such as SVMs and neural networks cannot obtain anchor points for a known-plaintext attack.
“thus further ensuring that no pattern-recognition algorithms, such as SVMs and neural-networks can be used to acquire anchor-points for a known-plaintext attack.”
Quote in the paperNot a substantial claim- Rests on an assumption: The chosen security metrics—histogram, correlation, information entropy, NPCR, and UACI—are sufficient to establish security against the considered attacks.
Classified as probably not a substantial claim.
PremisecitationConfusion and diffusion are essential properties for securing a cipher against analytical attacks.
“In order to secure any given cipher from any such analytical approaches, it is essential that two key properties called confusion and diffusion, first identified by Shannon (1949), be incorporated into it.”
Quote in the paperSubstantial claim0 of 1 sources checked- cites Shannon (1949) source
The paper relies on this without support.
The shared secret key remains secret from an attacker.
“via the usage of a shared secret key and a unique encryption/decryption vector”
The paper relies on this without support.
Sending the encryption vector EV with the ciphertext over an insecure channel does not let an attacker recover the plaintext or the secret key.
“C is finally transmitted on the insecure channel as the output of the entire process, along with EV, to be treated as the encrypted counterpart of input image P.”
The paper relies on this without support.
The chosen security metrics—histogram, correlation, information entropy, NPCR, and UACI—are sufficient to establish security against the considered attacks.
“The proposed algorithm is subjected to a number of theoretical, experimental, and mathematical analyses, to examine its performance and security against a number of possible attacks, using the following metrics - histogram analysis, correlation analysis, information entropy analysis, NPCR and UACI.”
The authors state this as an assumption.
The two parties have mutually agreed on a common secret initial key, and that key remains secret throughout the algorithm's operation.
“the algorithm assumes that the two parties involved in this operation, have mutually agreed upon a common value of IK. IK is assumed to be secret throughout the working span of the presented algorithm.”
The authors state this as an assumption.
The initialization vector is unique for every run of the algorithm.
“It is also essential that the value of IV is unique for each and every run”
The paper relies on this without support.
For a truly random image, the expected NPCR and UACI values are 99.6094% and 33.4653%, respectively.
“It should be noted that for a truly random image, the expected NPCR and UACI values are 99.6094% and 33.4653%, respectively.”
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Decryption module output is described inconsistently
MinorThe decryption section says the wrapped computation returns the encrypted image, but the same passage later says the process returns the decrypted counterpart. Those two descriptions conflict about what the module actually outputs.
“The values computed for K, N, G, and MAX, along with C are used to place a call to the GPU_wrapper module, which returns the encrypted image C.”
“P is finally returned as the output of the entire process, to be treated as the decrypted counterpart of the received image”
N is defined as bytes but later called pixels
MinorThe PRNG definition makes N the required number of random bytes. The encryption steps instead describe N as the number of pixels requested from that PRNG call.
“N signifies the required number of random bytes”
“the number of pixels required N”
Cites
Communication Theory of Secrecy Systems
Shannon, C. E. (1949) · Bell System Technical Journal · doi:10.1002/j.1538-7305.1949.tb00928.x · Cited by 9,440 · cited 1× · see citing works
“In order to secure any given cipher from any such analytical approaches, it is essential that two key properties called confusion and diffusion, first identified by Shannon (1949), be incorporated into it.”
BackgroundSecure integration of IoT and Cloud Computing
Stergiou, C. L., Psannis, K. E., Kim, B., & Gupta, B. B. (2016) · Future Generation Computer Systems · doi:10.1016/j.future.2016.11.031 · Cited by 1,114 · cited 2× · see citing works
“( Boritz and No 2005; Stergiou et al. 2018; Kalra and Sood 2015) Particularly in the instances where the data being dealt with is of image/video domain, cryptography and security are one of the first concerns raised by the people, governments, and law authorities alike.”
BackgroundGenetic Algorithms and Genetic Programming
Affenzeller, M., Wagner, S., Winkler, S., et al. (2009) · doi:10.1201/9781420011326 · Cited by 495 · cited 2× · see citing works
“Affenzeller et al. (Affenzeller et al. 2009) describe the general architecture of genetic algorithms, which can be used for the purpose of optimization of a given function.”
BackgroundSecure authentication scheme for IoT and cloud servers
Kalra, S., & Sood, S. K. (2015) · Pervasive and Mobile Computing · doi:10.1016/j.pmcj.2015.08.001 · Cited by 289 · cited 2× · see citing works
“( Boritz and No 2005; Stergiou et al. 2018; Kalra and Sood 2015) Particularly in the instances where the data being dealt with is of image/video domain, cryptography and security are one of the first concerns raised by the people, governments, and law authorities alike.”
BackgroundA hybrid genetic algorithm and chaotic function model for image encryption
Abdullah, A. H., Enayatifar, R., & Lee, M. (2012) · AEU - International Journal of Electronics and Communications · doi:10.1016/j.aeue.2012.01.015 · Cited by 206 · cited 1× · see citing works
“(Abdullah et al. 2012) make use of the genetic algorithms in the process of image encryption process, via a chaotic function logistic map.”
MethodSecurity in XML-based financial reporting services on the Internet
Boritz, J. E., & No, W. G. (2005) · Journal of Accounting and Public Policy · doi:10.1016/j.jaccpubpol.2004.12.002 · Cited by 97 · cited 3× · see citing works
“(D’agapeyeff 2016; Davies 1997; Boritz and No 2005) Many applications and algorithms have been developed to address these needs, and this paper also follows suite.”
BackgroundGPU computing with NVIDIA CUDA
Buck, I. (2007) · doi:10.1145/1281500.1281647 · Cited by 91 · cited 1× · see citing works
“(Buck 2007; NVIDIA 2020; Zeller 2011).”
Random number generators for massively parallel simulations on GPU
Manssen, M., Weigel, M., & Hartmann, A. K. (2012) · The European Physical Journal Special Topics · doi:10.1140/epjst/e2012-01637-8 · Cited by 64 · cited 2× · see citing works
“Manssen et al. (Manssen et al. 2012) provide a review of the existing CUDA-based random-number generators, in the context of the massively parallel simulations, like the Monte Carlo and molecular dynamics simulations.”
MethodA brief history of cryptography
Davies, D. (1997) · Information Security Technical Report · doi:10.1016/s1363-4127(97)81323-4 · Cited by 62 · cited 2× · see citing works
“(D’agapeyeff 2016; Davies 1997; Boritz and No 2005) Many applications and algorithms have been developed to address these needs, and this paper also follows suite.”
BackgroundA fast image encryption scheme based on AES
Zhang, Y., Li, X., & Hou, W. (2017) · 2017 2nd International Conference on Image, Vision and Computing (ICIVC) · doi:10.1109/icivc.2017.7984631 · Cited by 44 · cited 15× · see citing works
“( Ramesh and Jain 2015; Bharadwaj and Sairabanu 2020) and (Zhang et al. 2017), are used for performing comparative analysis with the proposed algorithm, to obtain better insights regarding performance and security parameters.”
BackgroundMethodResultInfluentialImproving SIMD efficiency for parallel Monte Carlo light transport on the GPU
van Antwerpen, D. (2011) · Proceedings of the ACM SIGGRAPH Symposium on High Performance Graphics · doi:10.1145/2018323.2018330 · Cited by 40 · cited 2× · see citing works
“(Antwerpen 2011; Zafar et al. 2010) The paper also presents a genetic optimization exercise in detail, that was performed to enhance the security metrics of the PRNG, by the process of hyper-parameter tuning.”
BackgroundMethodRiek, M., & Böhme, R. (2018) · Journal of Cybersecurity · doi:10.1093/cybsec/tyy004 · Cited by 39 · cited 1× · see citing works
“(Altigani et al. 2019; Armin et al. 2016; Riek and Bo¨hme 2018).”
Background2020 Cybercrime Economic Costs: No Measure No Solution
Armin, J., Thompson, B., Ariu, D., et al. (2015) · doi:10.1109/ares.2015.56 · Cited by 36 · cited 1× · see citing works
“(Altigani et al. 2019; Armin et al. 2016; Riek and Bo¨hme 2018).”
BackgroundFast implementation of block ciphers and PRNGs in Maxwell GPU architecture
Lee, W., Cheong, H.-S., Phan, R. C. ‐., & Goi, B. (2016) · Cluster Computing · doi:10.1007/s10586-016-0536-2 · Cited by 32 · cited 2× · see citing works
“(Lee et al. 2016) examine the implementation of the block ciphers like AES, CAST, Blowfish, etc.”
MethodText Encryption Algorithms based on Pseudo Random Number Generator
Mishra, M., & Mankar, V. H. (2015) · International Journal of Computer Applications · doi:10.5120/19507-0756 · Cited by 27 · cited 2× · see citing works
“Mishra et al. (Mishra and Mankar 2015) propose algorithms for the encryption and decryption of textual data, using pseudo-random generator, and linear congruential generator.”
Method(2010) · doi:10.5555/1921479.1921500 · Cited by 27 · see citing works
Cybercrime Economic Costs: No Measure No Solution
Armin, J., Thompson, B., & Kijewski, P. (2016) · Advanced sciences and technologies for security applications · doi:10.1007/978-3-319-38930-1_8 · Cited by 26 · see citing works
Combatting Cybercrime and Cyberterrorism
Akhgar, B., & Brewster, B. (2016) · Advanced sciences and technologies for security applications · doi:10.1007/978-3-319-38930-1 · Cited by 24 · see citing works
Dutta, S., Jash, S., Patra, D., et al. (2014) · Cited by 19 · see citing works
Ramesh, A., & Jain, A. (2015) · 2015 International Conference on Trends in Automation, Communications and Computing Technology (I-TACT-15) · doi:10.1109/itact.2015.7492652 · Cited by 19 · cited 6× · see citing works
“( Ramesh and Jain 2015; Bharadwaj and Sairabanu 2020) and (Zhang et al. 2017), are used for performing comparative analysis with the proposed algorithm, to obtain better insights regarding performance and security parameters.”
BackgroundMethodInfluentialAlmalkawi, I. T., Halloush, R. D., Alsarhan, A., et al. (2019) · Journal of Information Security and Applications · doi:10.1016/j.jisa.2019.102384 · Cited by 17 · cited 2× · see citing works
“(Almalkawi et al. 2019) present an image encryption algorithm, that is designed to be useful for the wireless network applications, that require the PRNG to be lightweight and efficient.”
MethodKösemen, C., Dalkılıç, G., & Aydın, Ö. (2018) · TURKISH JOURNAL OF ELECTRICAL ENGINEERING & COMPUTER SCIENCES · doi:10.3906/elk-1710-155 · Cited by 12 · see citing works
Altigani, A., Hasan, S., Shamsuddin, S. M., et al. (2019) · Journal of Information Security and Applications · doi:10.1016/j.jisa.2019.03.013 · Cited by 10 · cited 1× · see citing works
“(Altigani et al. 2019; Armin et al. 2016; Riek and Bo¨hme 2018).”
BackgroundSeshadrinathan, M., & Dempski, K. L. (2008) · 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops · doi:10.1109/cvprw.2008.4563094 · Cited by 9 · cited 2× · see citing works
“Table 5 compares the average encryption and decryption speeds of the proposed algorithm, and the ones proposed in Failed (2015); Bharadwaj and Sairabanu 2020; Zhang et al. 2017) and (Seshadrinathan and Dempski 2008) respectively.”
MethodNon-standard pseudo random number generators revisited for GPUs
Riesinger, C., Neckel, T., & Rupp, F. (2016) · Future Generation Computer Systems · doi:10.1016/j.future.2016.12.018 · Cited by 9 · cited 2× · see citing works
“Riesinger et al. (Riesinger et al. 2018) review the non-standard PRNGs for normally distributed random numbers in the context of GPU implementation.”
BackgroundBit level symmetric key cryptography using genetic algorithm
Sen, A., Ghosh, A., & Nath, A. (2017) · doi:10.1109/csnt.2017.8418536 · Cited by 9 · cited 2× · see citing works
“Table 5 compares the average encryption and decryption speeds of the proposed algorithm, and the ones proposed in Failed (2015); Bharadwaj and Sairabanu 2020; Zhang et al. 2017) and (Seshadrinathan and Dempski 2008) respectively.”
MethodCodes and Ciphers - A History Of Cryptography
d'Agapeyeff, A. (2015) · Cited by 9 · see citing works
Parallel implementation of cryptographic algorithm: AES using OpenCL on GPUs
Inampudi, G. R., Shyamala, K., & Ramachandram, S. (2018) · 2018 2nd International Conference on Inventive Systems and Control (ICISC) · doi:10.1109/icisc.2018.8398949 · Cited by 5 · cited 3× · see citing works
“Inampudi et al. (Inampudi et al. 2018) present an implementation of the AES algorithm on the GPGPU, using the OpenCL API, via the usage of data decomposition technique, where they divide the workload between the 256 concurrently executing threads of the GPU, and compare their implementation to the serial version of th…”
MethodInfluentialChaotic image encryption on <i>GPU</i>
Vihari, P. L. V., & Mishra, M. (2012) · Proceedings of the CUBE International Information Technology Conference · doi:10.1145/2381716.2381860 · Cited by 4 · cited 1× · see citing works
“(Antwerpen 2011; Zafar et al. 2010) The paper also presents a genetic optimization exercise in detail, that was performed to enhance the security metrics of the PRNG, by the process of hyper-parameter tuning.”
BackgroundImage Encryption using a Modified Pseudo-Random Generator
Bharadwaj, B., & Sairabanu, J. (2020) · 2020 International Conference on Emerging Trends in Information Technology and Engineering (ic-ETITE) · doi:10.1109/ic-etite47903.2020.094 · Cited by 4 · cited 12× · see citing works
“( Ramesh and Jain 2015; Bharadwaj and Sairabanu 2020) and (Zhang et al. 2017), are used for performing comparative analysis with the proposed algorithm, to obtain better insights regarding performance and security parameters.”
MethodResultInfluentialSelf-citationAn Image Encryption Scheme Based on Fractal Interpolation
Ye, R., Li, Y., & Li, Y. (2018) · International Conference on Multimedia and Image Processing (ICMIP) · doi:10.1145/3195588.3195596 · Cited by 3 · cited 2× · see citing works
“(Ye et al. 2018).”
Tiny encryption algorithm for parallel random numbers on the GPU
Zafar, F., & Olano, M. (2010) · Proceedings of the 2010 ACM SIGGRAPH symposium on Interactive 3D Graphics and Games · doi:10.1145/1730804.1730973
Flagged first, then by citations
Research works printed but not linked: 2
- Bharadwaj B, Shukla S, Shalini L (2018) Symmetric key encryption using a simple pseudo-random generator to provide more secure communication. Int J Tech Res Appl 6(2):76–81
- Zeller C (2011) Cuda c/c++ basics. NVIDIA Coporation, California
Cited by
Sorted by citations
Applications of graph’s complete degree with bipolar fuzzy information
Poulik, S., & Ghorai, G. (2021) · Complex & Intelligent Systems · doi:10.1007/s40747-021-00580-x · Cited by 32 · see citing works
PixAdapt: A novel approach to adaptive image encryption
Tuli, R., Soneji, H. N., & Churi, P. P. (2022) · Chaos Solitons & Fractals · doi:10.1016/j.chaos.2022.112628 · Cited by 26 · see citing works
Kıran, H. E. (2024) · Chaos and Fractals · doi:10.69882/adba.chf.2024072 · Cited by 25 · see citing works
Ullah, H., Ahmad, B., Sana, I., et al. (2021) · CAAI Transactions on Intelligence Technology · doi:10.1049/cit2.12046 · Cited by 19 · see citing works
ECC based novel color image encryption methodology using primitive polynomial
Sharma, P., Gupta, S., Nayyar, A., Harish, M., Gupta, K., & Sharma, A. K. (2024) · Multimedia Tools and Applications · doi:10.1007/s11042-024-18245-5 · Cited by 13 · see citing works
“Chaotic systems exhibit complex and unpredictable behavior, making them suitable for enhancing the security of image encryption [5–16].”
Performance evaluation of spatial fuzzy C-means clustering algorithm on GPU for image segmentation
Ali, N. A., Abbassi, A. E., & Bouattane, O. (2022) · Multimedia Tools and Applications · doi:10.1007/s11042-022-13635-z · Cited by 9 · see citing works
“…and signal processing applications, and heuristic optimization method such as (Smith-Waterman algorithm, genetic quantum algorithm, simulated annealing, particle swarm and the solution of differential equation) are few examples that have been implemented successfully using CUDA [4, 12, 37, 53].”
MethodFinger vein recognition method based on ant colony optimization and improved EfficientNetV2
Ma, X., & Luo, X. (2023) · Mathematical Biosciences & Engineering · doi:10.3934/mbe.2023490 · Cited by 8 · see citing works
A genetic algorithm for subcontractors selection and allocation in multiple building projects
Afshar, M. R., Shahhosseini, V., & Sebt, M. H. (2021) · Soft Computing · doi:10.1007/s00500-021-06008-5 · Cited by 6 · see citing works
Jadhav, S. V., Patel, U. S., Natu, A., et al. (2023) · Lecture notes on data engineering and communications technologies · doi:10.1007/978-981-99-1767-9_23 · Cited by 3 · see citing works
Amutha, R., & Phamila, Y. A. V. (2026) · Scientific Reports · doi:10.1038/s41598-026-40337-5 · Cited by 2 · see citing works
Aldin, S. S. A. B., Aldin, N. B., & Aykaç, M. (2026) · Engineering Science and Technology an International Journal · doi:10.1016/j.jestch.2026.102287
Баранов, И. В. (2025) · Lecture notes in networks and systems · doi:10.1007/978-3-031-92611-2_20
A Multiple Precision Floating-Point Arithmetic Unit Based on the RISC-V Instruction Set
Chen, J., Hao, H., Wang, S., et al. (2024) · doi:10.1109/eiect64462.2024.10867213
“Graphics Processing Units (GPUs) [1] have become one of the mainstream high-performance computing platforms due to their high concurrency and good programmability.”
- Authors
- Bharadwaj, B., Saira Banu, J., Madiajagan, M., Ghalib, M. R., Castillo, O., & Shankar, A.
- Affiliations
- Vellore Institute of Technology University (IN)
- Instituto Tecnológico de Tijuana (MX)
- Amity University (IN)
- Corresponding author
- Shankar, Achyut
- Venue
- Soft Computing
- Published
- 2021-09-30
- Volume
- 25
- Issue
- 22
- Pages
- 14413-14428
- ISSN
- 1432-7643 (ISSN-L)
- 1433-7479
- PMC
- PMC8483430
- OpenAlex
- W3201876383
- Semantic Scholar
- 736172dbaa1f9f3ea76cb2788e40e562b24f1d20
- Cited by
- 14
- References
- 31
- Citations by year
- 2026: 3 · 2025: 1 · 2024: 3 · 2023: 2 · 2022: 2 · 2021: 3
- FWCI
- 0.90
- Type
- journal article
- Language
- en
- Open access
- bronze
- doi.org
- Licence
- Locations
- Soft Computing (publishedVersion)
- PubMed Central (publishedVersion)
- PubMed (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Chaos-based Image/Signal Encryption
- Physical Sciences › Computer Science › Artificial Intelligence › Computer Science and Engineering
- Physical Sciences › Computer Science › Artificial Intelligence › Cryptographic Implementations and Security
- Keywords
- image encryption, GPU acceleration, genetic algorithm optimization, XOR cipher, pseudorandom number generators, parallel encryption, NPCR, UACI, Shannon entropy, histogram analysis, correlation analysis
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Claims read under definition version 3 (extract-v3) on 2026-09-28.
Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-09-28.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
