Crayfish optimization based pixel selection using block scrambling based encryption for secure cloud computing environment (2025)

Cloud storage is a main part of a cloud computing (CC) method that is used to save information and useful data over the internet¹. The Cloud providers offer users using cloud storage service at a very low cost as per the need and efficiently manage data storage structure. With the quick growth of cloud storage technology, enterprises, single users, and governments progressively retain data in the cloud². Though CC delivers a huge number of services such as the act of outsourcing user data safety to avoid unauthorized consumers, the costs of source maintenance, the confidentiality of sensitive information, and computational difficulty are the main complications. With the fast growth of information technology (IT), the security of multi-media data like audio, video, and image has attracted extensive consideration³. Safe and effective encryption of image information is the main focus of many multi-media research. Images are the most used method of multi-media objects for a wide range of applications. In a cloud platform, image security is a major problem to be mentioned⁴. Images are encoded both in transform and spatial domains based on different parameters and requirements to increase privacy. Owing to the correlation dependence of image pixels, conventional cryptographic methods are not suitable for ciphering the images⁵. Owing to the small entropy of digital images, including strong pixel correlation and high redundancy, traditional encryption techniques usually could not efficiently encrypt the image data. The novel cryptographic algorithm based on a chaotic system has become one of the best image encryption solutions⁶.

Cryptography is a method, which includes the research of safe communication methods to prevent unauthorized persons from retrieving confidential data, messages, or information. The method includes several characteristics in the information security sector, namely data confidentiality, integrity, authentication (CIA), and non-repudiation, which is the main area in current cryptography⁷. Encryption in CC is a significant problem that requires investigation in several studies. Because CC manages critical data and is available worldwide over the internet, security is a serious issue and major concern. Cryptography techniques play an important part in protecting digital media privacy, data transmission, e-commerce, web data storage, and transmission⁸. Some procedures are suitable for decryption and encryption to confirm data safety in CC including Data Encryption Standard (DES), Identity-Based Encryption (IBE), Advanced Encryption Standard (AES), and Rivest Shamir Adleman Algorithm (RSA). Many researchers utilize cryptography methods for protecting the cloud privacy of the data, but the main drawback of encryption is that the data is encrypted and turned unreadable, still occurring as private data⁹. If the hacker has enough time, he/she can decrypt the private data. Steganography is a method to resolve this problem because it will permit the users to disappear the data into other objectives like images, text, audio, and video¹⁰. These methods will upsurge the security of confidential data.

In the field of cloud computing, securing sensitive data, particularly digital images, is increasingly critical. Traditional cryptographic and steganographic methods often fall short in providing comprehensive security, necessitating advanced approaches like the Crayfish Optimization-based Pixel Selection with Block Scrambling Based Encryption Approach (CFOPS-BSBEA). Traditional cryptographic techniques, such as AES, are effective for encrypting data, but when combined with basic steganographic methods, they can still be vulnerable to sophisticated attacks. For example, simple steganographic methods that hide data in the least significant bits (LSBs) of pixels can be detected by advanced forensic tools. This limitation is evident in scenarios such as secure medical image transmission, where maintaining data confidentiality and integrity is crucial. Basic steganography fails to protect high-resolution images effectively against detection and extraction attempts. Moreover, many conventional stego-crypto methods do not optimize pixel selection for embedding secret data, leading to predictable patterns that can be exploited by attackers. For instance, in the secure transmission of satellite imagery, using fixed pixel locations or predictable encryption keys can increase the risk of unauthorized access. The lack of optimization and randomness in these methods makes them less effective in high-security environments. The CFOPS-BSBEA approach addresses these issues by incorporating Crayfish Optimization to dynamically select pixels for embedding secret data, thus enhancing security through optimized pixel selection. This is particularly beneficial in secure military communications, where the ability to prevent unauthorized detection of hidden information is crucial. Additionally, the approach utilizes block scrambling encryption, which adds a layer of complexity that traditional methods lack. This technique disrupts predictable patterns, significantly improving security. For example, in the secure transfer of financial documents, block scrambling makes it much harder for unauthorized parties to uncover hidden data. Overall, the CFOPS-BSBEA approach offers a robust solution to the limitations of traditional methods, providing a higher level of security in cloud computing environments.

This study develops a new Crayfish Optimization based Pixel Selection with Block Scrambling Based Encryption Approach (CFOPS-BSBEA) technique for Secure CC Environment. In the CFOPS-BSBEA technique, a three-stage process is involved. Firstly, the CFOPS-BSBEA technique involves the design of a CFO algorithm for optimally selecting the pixels for the steganography process. Next, the CFOPS-BSBEA technique applies the BSBE technique to encrypt the secret images that are embedded in the selected pixels of the coverd image. The keys involved in the BSBE process can be selected by the use of the Salp swarm algorithm (SSA). At last Finally, the CFOPS-BSBEA technique undergoes the embedding and extraction process. The experimental results highlighted that the CFOPS-BSBEA technique reaches better security performance than other models.

In¹¹, a new hybrid technique, CNN-DCT Steganography unites the influence of CNN and the discrete cosine transform (DCT) technique has been developed. The projected model exploits the strong feature extractor abilities of CNN and the spatial frequency field alteration of DCT for attaining unnoticeable embedding and improved data-hiding ability. The cover image experiences a dual-stage procedure. Initially, feature extraction utilizing a deep CNN that allows many suitable areas for data embedding. Then, the nominated areas are exposed to the DCT-based steganography model. Bahaddad et al.¹² concentrated on the project of Bald Eagle Search Optimum Pixel Selection with Chaotic Encryption based on the image steganography model. The projected Chaotic Encryption (BESOPS-CE) method efficiently covers the secret imagery in its encoded form for covering an image. To attain it, the BESOPS-CE model uses a BES for the OPS process. Also, c Chaotic encryption has also been implemented for encoding the image of secret, then it is fixed to select the pixel point of the hidden imagery. Lastly, embedding and extraction procedures have been implemented. Sharath et al.¹³ proposed an optimum meta-heuristics-based PS with a homomorphic encryption model for the video steganography (OMPS-HEVS) method. The projected OMPS-HEVS system at first achieves a frame conversion procedure and relates a 2D-DWT procedure. Also, the OPS procedure utilizes the glowworm swarm optimizer (GSO) model.

Guo et al.¹⁴ projected an Adaptable Image Steganography Method (AISM). At first, AISM modifies correct down-sampling models, ratios, and up-sampling ratios for secret images depending upon the desires, tracked by the image sample procedure. By following this, an embedding technique based on pixel-value coding is planned, which plans pixel values and then substitutes the higher-frequency sub-band coefficient of the up-sampling image. In the embedding procedure, no auxiliary data is produced, which is a key feature for user-friendliness. Ren et al.¹⁵ presented a visual refuge image encryption structure united with compacted detecting and LSB embedding in the cloud atmosphere. Initially, the Arnold technique was employed to challenge the sparse plain-text imagery on the local consumer, and compacted sensing was employed for compressing it to get the secret imagery. Upload the cipher-text image and implement encryption in the cloud. The security of visuals is certified by inserting cipher text into carrier imageries to get significant steganographic imageries. In¹⁶, the developed model has concentrated on executing the least significant bit (LSB) identical steganography method. For certifying better safety, the advanced encryption standard (AES) system has been employed before using the steganography method to certify the dual-layer safety of the private message. In this study, another feature has been employed and uses mosaic imageries as the cover media.

The Crayfish Optimization-based Pixel Selection with Block Scrambling Based Encryption Approach (CFOPS-BSBEA) represents a significant advancement in image security, particularly in the context of cloud computing. Traditional stego-crypto methods often face limitations in balancing capacity and secrecy. For example, researches in^23,24 highlights how conventional methods can struggle with achieving both high data capacity and robust security. CFOPS-BSBEA overcomes these challenges by employing Crayfish Optimization to enhance pixel selection, which optimizes the embedding process and improves security without compromising the capacity to hide data. This approach is further supported by findings in²⁵ demonstrate that advanced techniques can significantly enhance security and performance. Additionally, CFOPS-BSBEA addresses the limitations of traditional methods highlighted in studies like paper in²⁶ by introducing a complex encryption process that makes unauthorized detection more difficult. Overall, CFOPS-BSBEA builds on existing research to provide a sophisticated solution that achieves superior security and efficiency in image data protection.

Kanjanamek et al.¹⁷ presented a cloud-based steganography method that provides imperceptibility protection and adaptive matching among files, text, or ciphertext within nominated cover imageries at random. To conclude, this method projected a method to calculate the optimum cover image for dissimilar cipher-text dimensions based on the enhanced calculation of the LSB and file ratio. Also, the model suggests a graph-based method to physically perfect the correlation among consumers, ciphertext, and image files. Alomoush et al.¹⁸ developed an unseen watermarking model for data embedding into the transformation field for the grey scale imageries. Likewise, a stego-text is inserted with dissimilar dimensions within imageries after implanting the stego-image immune to dissimilar types of attacks like cropping, rotation, JPEG compression, and salt and pepper with diverse standards.

Our proposed CFOPS-BSBEA technique, which combines steganography and encryption, shares similarities with other studies that have explored the integration of cryptography and steganography to enhance security in various applications. For instance, the study²⁷ demonstrates the effectiveness of combining elliptic curve cryptography with image steganography to secure medical data. Similarly, the study²⁸ proposes a 3-layer security approach that incorporates cryptography and steganography to protect medical records. Another study²⁹, also explores the combination of elliptic curve cryptography and image steganography to secure medical data.

In our study, we have designed a novel CFOPS-BSBEA method that incorporates steganography and encryption to improve security in cloud computing environments. The CFOPS-BSBEA technique involves a three-stage procedure, which includes pixel selection using the Crayfish Optimization (CFO) algorithm, encoding secret images using Block Scrambling Based Encryption, and embedding them in cover images. The CFO algorithm is a new swarm intelligence based optimizer algorithm inspired by the behavior of crayfish, which is used to optimally select pixels for the steganography process. Our approach offers a unique solution to improve security in cloud environments, and its originality and novelty make it an attractive contribution to the field. While the referenced studies by^30,31,32 focus on improving encryption confidentiality through various means of randomness and resilience, the CFOPS-BSBEA method offers a more integrated approach. By combining Crayfish Optimization with Block Scrambling Based Encryption, CFOPS-BSBEA provides a holistic solution that enhances both the randomness and structural security of image encryption. This multifaceted approach makes it a robust option for high-resilience applications in cloud computing environments. For instance, the study by Alghamdi et al. (2022) explores securing matrix counting-based secret-sharing through crypto steganography, highlighting how these techniques can bolster data protection against unauthorized access³³. Similarly, the work by Ahmad et al. (2022) demonstrates the application of counting-based secret sharing for lightweight semi-complete authentication in watermarking images, showcasing the utility of combining these methods for improved security³⁴. Furthermore, the research conducted by Shaikh et al. (2022) discusses increasing participant capacity in counting-based secret sharing by incorporating matrices and practical steganography, emphasizing the scalability of these approaches³⁵. Lastly, the paper by Syed et al. (2021) refines image steganography distribution for enhanced security in multimedia through counting-based secret-sharing, reflecting on the benefits of integrating these strategies³⁶.

In this study, A new we design of a novel CFOPS-BSBEA method for a secure CC environment is introduced. The CFOPS-BSBEA technique offers a novel approach to improve security in the cloud environment by the incorporation of the steganography and encryption scheme³⁷. In the CFOPS-BSBEA system, a three-stage procedure is demonstrated in Fig.1.

Pixel selection using CFO algorithm

Firstly, the CFOPS-BSBEA technique involves the design of a CFO algorithm for optimally selecting the pixels for the steganography process. CFO is a new swarm intellect optimizer algorithm stimulated by crayfish in summer temperature, predation behavior, and competition¹⁹. Crayfish are arthropods that belong to the family of shrimp. It mostly lives in freshwater regions. Research has displayed that crayfish can perform differently in diverse ambient temperatures. In the numerical modeling of CFO, the competition, heat escape, and behavior of predation are definite as 3 different phases, and the optimizer algorithm is measured to arrive at diverse phases by describing diverse temperature ranges. Amongst them, the summer phase is the exploration phase, and the foraging and competition phase is the development phase of the CFO. The stages of CFO are defined below in detail.

Initialization of population.

The CFO is a population-based method, which begins with the initialization of the population to deliver an appropriate initial point for the next optimizer procedure. In the demonstration of CFO, every crayfish position signifies a candidate solution to an issue that contains D dimension, and the population position of \(\:N\) crayfish establishes a set of candidate solution \(\:X\), whose matrix is exposed in Eq.(1).

Here, \(\:X\) denotes the position of the early population of crayfish, \(\:N\) refers to the no. of crayfish population, \(\:D\) represents the size of the issue and \(\:{X}_{i,j}\) is the preliminary position of \(\:i\) crayfish in the \(\:j\) dimension that is produced in the search space at random, and the specific formulation of \(\:{X}_{i,j}\) is exposed in Eq.(2).

Whereas, \(\:l{b}_{j}\) and \(\:u{b}_{j}\) denote the lower and upper bound of the \(\:jth\:\)dimension, respectively; \(\:r\) refers to the evenly distributed arbitrary number, which belongs to \(\:\left[\text{0,1}\right].\)

Temperature and Crawfish Food Consumption.

At dissimilar ambient temperatures, the crayfish will arrive in dissimilar phases. It will go in the summer phase if the \(\:Temp\) is greater than \(\:3{0}^{o}C\). Crayfish have aggressive behavior of predation among \(\:1{5}^{o}C\) and \(\:3{0}^{o}C\), with \(\:2{5}^{o}C\) being the optimum temperature³⁸. Also, their food consumption is assumed by temperature and it nearly happens when temperature varies. In CFO, the \(\:Temp\:\)is definite in Eq.(3).

Whereas, \(\:Temp\:\)is the temperature of ambient. The numerical calculation of food consumption \(\:P\) of crayfish is exposed in Eq.(4).

Here, \(\:\mu\:\) denotes the optimum temperature; \(\:{C}_{1}\) and \(\:\sigma\:\) are employed to switch the food consumption of crayfish at dissimilar temperatures of ambient.

Summer Phase.

If the \(\:Temp\) is greater than 30C, then the crayfish will select \(\:{X}_{shade}\) cavern for heat escape, which is called as heat escape phase of CFO. The calculation of \(\:{X}_{shade}\) cavern is revealed in Eq.(5).

Whereas, \(\:{X}_{G}\) denotes the optimum location attained by the method iteration until now, and \(\:{X}_{L}\) represents the optimum location of the present crayfish population.

There will struggle for crawfish to acquire into the hotness. Many crayfish will participate in a similar hole to escape the temperature if there are numerous crayfish and fewer burrows. If there were many caves, this would not be the situation. A randomly produced integer among \(\:0\) and 1, \(\:rand\:\)is employed to define whether competition has happened in CFO. If the randomly generated integer of \(\:rand<0.5\), no other crayfish participate in the cavern, and crayfish can straight go into the cavern to escape from the heat³⁹. The numerical formulation of this procedure is presented in Eq.(6).

Here, \(\:r\) denotes the present iteration count, \(\:{X}_{i,j}^{t}\) refers to the present site of the \(\:i\) crayfish, \(\:t+1\) signifies the iterations count of the next group, \(\:r\) denotes a randomly generated integer [0, 1], and \(\:{C}_{2}\) values reduce with the upsurge in iterations, as stated in Eq.(7).

Whereas \(\:T\) denotes the maximum iteration count.

Competition Stage.

Many crayfish will contest for a cavern and arrive at the opposition phase if the \(\:Temp\) is greater than \(\:3{0}^{o}C\) and the randomly produced integer \(\:rand\ge\:0.5\). In this phase, the crayfish location is upgraded, which is exposed in Eq.(8).

Whereas, \(\:z\) is an arbitrary crayfish, and its formulation is revealed in Eq.(9).

Here, \(\:r\) denotes the randomly produced integer, which belongs to [0 and 1], and \(\:round\) refers to the number function.

Predation Phase.

The crayfish will search for and consume food if the \(\:Temp\le\:3{0}^{o}C\). The Crayfish travel near their nutrition and consume it. The position of food \(\:{X}_{food}\) is definite in Eq.(10).

The crayfish will estimate the food size to assume dissimilar methods before feeding nutrition. The dimension \(\:Q\) of food is definite in Eq.(11). The crayfish use to open the food using their nails at primary. When the food size is very big, they intake with their 2nd and 3rd walking feet.

Whereas, \(\:{C}_{3}\) denotes the food feature, demonstrating the highest value of food; \(\:Fitnes{s}_{i}\) represents the fitness value of \(\:i\) crayfish; \(\:Fitnes{s}_{food}\) signifies the fitness value of food position \(\:{X}_{food}\). The Crayfish evaluate the food size by \(\:{C}_{3}\) of their greatest nutrition. If the food dimension is \(\:Q>({C}_{3}+1)/2\), then the food is very big, and the small dragon will utilize chelates (shrimp nails) to open the food. The numerical calculation is shown in Eq.(12).

Next, the crayfish will substitute food with the 2nd and 3rd feet, a procedure pretend in the CFO utilizing cosine and sine \(\:functSin2\) in Eq.(13).

Whereas, \(\:P\) refers to the food consumption and \(\:r\) denotes the randomly produced integer, which belongs to [0 and 1].

If the dimension of food is suitable, then the crayfish can straight eat when \(\:Q\le\:({C}_{3}+1)/2\), and the location upgrade calculation was revealed in Eq.(14).

Here, \(\:r\) is a randomly produced number that belongs to \(\:\left[\text{0,1}\right].\) Algorithm 1 depicts the steps involved in CFO.

The fitness function (FF) used in the CFO technique is intended to consume a balance between numerous nominated features in every solution (minimum) and the classification accuracy (maximum) attained by utilizing these nominated features, Eq.(15) signifies the FF to assess solution⁴⁰.

Here, \(\:{\gamma\:}_{R}\left(D\right)\) signifies the classifier rate of error of an assumed classifier. \(\:\left|R\right|\:\)represents the cardinality of the nominated sub-set and \(\:\left|C\right|\) refers to the total sum of features in the dataset, \(\:\alpha\:\) and \(\:\beta\:\) refer to the dual parameters equivalent to the consequence of classifier excellence and sub-set length. ∈ [1,0] and \(\:\beta\:=1-\alpha\:.\)

Image encryption using BSBE technique

Next, the CFOPS-BSBEA technique applies the BSBE technique to encode the secret images which are afterward embedded in the selected pixels of the cover imagery. The BSBE model is used for security²⁰. In this method, a user wants to firmly convey the image \(\:I\) to viewers utilizing providers of Social Networking Services (SNS). If the consumer cannot deliver the confidential key \(\:K\) to SNS providers, image privacy is revealed beneath the control of customers even if the SNS provider re-compresses the image \(\:I\). Hence, the user is made secure of confidentiality by themselves. When equated to the \(\:CtE\) method, the user is to reveal unencrypt imageries to re-compress them.

In this model, an image with \(\:X\times\:Y\) pixel is mainly separated as a non-overlapped block with \(\:{B}_{x}\times\:{B}_{y}\); next, 4 blocks of scrambling‐based processing phases were implemented for dividing the images. The steps to perform encryption of an image for making an encryption image \(\:{I}_{e}\) is mentioned in algorithm2.

SSA-based key selection process

At this stage, the keys involved in the BSBE process can be selected by the use of SSA. Mirjalili et al. presented the SSA as a new member of the swarm optimizer algorithms group²¹. The main objective of SSA is to imitate the combined features of salps. During the aquatic presence, salps are involved in an individual swarming habit related to the “salp chain”, which is also deployed in its search for food. Further,the population of SSA comprises 2 different groups such as followers and leaders. The leader in the salp chain plays a critical role in determining movement, feeding locations, and sometimes updating these target spot selections. The remaining individuals are designated by “followers,” and everyone subsequently obeys the leader in sequence, creating a chain design. Each single point from the \(\:n\)-dimensional searching region defines the potential result with \(\:n\) demonstrating the count of variables appropriate to problems. Furthermore, the model of “food supply” represented by \(\:F\) implies the objective function.

\(\:{x}_{i}^{1}\) and \(\:F\) signify the leader and target place from the \(\:{j}^{th}\) dimensional, whereas \(\:u{b}_{j}\) and \(\:l{b}_{j}\) signify the upper and lower bounds, respectively. \(\:r2\) and \(\:r3\) scalar rates are arbitrarily selected from the range of zero and one. The vital control parameter is \(\:r1\), in control of stabilizing the exploitation and exploration. \(\:r1\) is written in Eq. \(\:(17\)):

At this point, the present iteration counts and the maximal potential iteration counts are referred by \(\:t\) and \(\:T\), correspondingly. The formula given in Eq. \(\:\left(18\right)\) was utilized to upgrade the followers’ locations so \(\:i>2.\)

Afterward, as realized in Eq.(20), Newton’s theory of motion can be utilized:

At this point, \(\:{x}_{j}^{i}\) refers to the \(\:{i}^{th}\) follower’s position from the \(\:{j}^{th}\) dimensional, \(\:t\) implies the iteration, \(\:{s}_{0}\) stands for the beginning speed, and \(\:k\) is expressed in Eq.(20):

With \(\:{s}_{0}\) described in Eq.(21)

Embedding and extraction process

At last, the CFOPS-BSBEA technique undergoes the embedding and extraction process. For an assumed image of cover \(\:C\), the encoded confidential transmission ES has been hidden to conceal the imagery. Initially, an IWT procedure was used to alter \(\:C\) in domains of spatial to frequency. The alteration results were separated into 4 sub-groups namely low‐high (LH), high‐high (HH), low‐low (LL), and high‐low (HL). Where ES is embedded from sub‐blocks of LH, HH, and HL as bits from \(\:k\)‐\(\:LSB\) of every pixel. After concealing every confidential information from the precise sub‐groups, it was served into OPAP to diminish the differences between the new and changed co-efficient. Next, the inverse transformation procedure was exposed to combine the sub‐groups and generate stego imagery. To remove the transmission of confidentiality, the stego imagery was transformed into a field of frequency by utilizing IWT. Next, the LSB was removed in every pixel of \(\:HH\),\(\:\:LH\), and \(\:HL\:\)sub‐groups. The isolated bit is the encoding procedure of ES. It used decoding exclusion bit with encoding vector \(\:Ev\) to get new bits. Lastly, the bits were changed to attain the confidential image \(\:S\).

This section analyzes the encryption results offered by the CFOPS-BSBEA technique. The proposed CFOPS-BSBEA technique can be examined using the color images from the USC-SIPI database²². In evaluating the CFOPS-BSBEA technique, a detailed analysis of payload expansion and error rates is crucial for understanding its security effectiveness. The technique’s payload expansion results from incorporating encryption into steganography, which increases the data volume and may affect the cover image’s quality. This expansion is quantified by metrics such as Mean Squared Error (MSE) and Root Mean Square Error (RMSE), which reflect the distortions introduced in the encrypted images. Lower MSE and RMSE values indicate minimal impact on image quality and effective data embedding. The connection between cover and stego images, assessed through Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index (SSIM), helps identify any anomalies that could signal potential security breaches. Furthermore, the encryption key’s robustness is critical in reducing the cracking probability; stronger keys significantly enhance security against attacks. By comparing the CFOPS-BSBEA method with existing models, it is evident that this technique performs favorably in terms of both payload efficiency and error rates, demonstrating its capability to maintain high security while minimizing detectable distortions. This comprehensive analysis ensures that the system not only achieves high encryption quality but also mitigates risks associated with data cracking and steganalysis. Figure2 represents the sample images.

Table1 presents the overall results obtained by the CFOPS-BSBEA method under several secret images (SIs). The results indicate that the CFOPS-BSBEA technique reaches effectual encryption results. With SI of IMG 1, the CFOPS-BSBEA technique provides MSE of 0.1680, RMSE of 0.4099, PSNR of 55.8777dB, and SSIM of 0.9996. Also, With SI of IMG 2, the CFOPS-BSBEA model delivers MSE of 0.1437, RMSE of 0.3791, PSNR of 56.5562dB, and SSIM of 0.9990. Besides, With SI of IMG 3, the CFOPS-BSBEA approach provides MSE of 0.1646, RMSE of 0.4057, PSNR of 55.9665dB, and SSIM of 0.9991. Finally, With SI of IMG 4, the CFOPS-BSBEA technique provides MSE of 0.1438, RMSE of 0.3792, PSNR of 56.5532dB, and SSIM of 0.9990.

Table2 portrays the overall computation time (CT) results of the CFOPS-BSBEA technique. The resultsshowed that the CFOPS-BSBEA technique properly encrypted the images with minimal CT. In Fig.3, the embedding time (EMBT) results of the CFOPS-BSBEA method is exhibited under numerous SEs. The figure shows that the CFOPS-BSBEA approach reaches reduced EMBT values. On SI of IMG 1, the CFOPS-BSBEA technique offers the EMBT of 0.4123min. Also, On SI of IMG 2, the CFOPS-BSBEA model offers the EMBT of 0.4183min. Meanwhile, On SI of IMG 3, the CFOPS-BSBEA approach gets the EMBT of 0.4105min.

In Fig.4, the extraction time (ET) results of the CFOPS-BSBEA system is shown under various SEs. The figure displayed that the CFOPS-BSBEA model gets reduced ET values. On SI of IMG 1, the CFOPS-BSBEA model offers the ET of 0.3136min. Moreover, On SI of IMG 2, the CFOPS-BSBEA method offers the ET of 0.3080min. While, On SI of IMG 3, the CFOPS-BSBEA approach offers the ET of 0.3397min.

In Table3, a comparative MSE and RMSE results of the CFOPS-BSBEA technique is demonstrated¹². Figure5 highlights the MSE results of the CFOPS-BSBEA technique in comparison with existing models. The figure stated that the CSOSM method has shown the least performance with maximum MSE values. In line with, the CEGAN, GSOSM, and SSOSM models have managed to obtain slightly reduced and closer MSE values. Furthermore, the BESOPSCE model has tried to accomplish considerable MSE values. Nevertheless, the CFOPS-BSBEA technique gains better performance with the least MSE of 0.1680, 0.1437, 0.1646, 0.1438, 0.1245, and 0.1332, correspondingly.

Figure6 highlights the RMSE outcomes of the CFOPS-BSBEA technique with present models. The figure identified that the CSOSM method has revealed minimum performance with maximum RMSE values. Whereas, the CEGAN, GSOSM, and SSOSM approaches have managed to find slightly reduced and closer RMSE values. Also, the BESOPSCE technique has tried to achieve considerable RMSE values. However, the CFOPS-BSBEA model gains enhanced performance with minimum RMSE of 0.4099, 0.3791, 0.4057, 0.3792, 0.3528, and 0.3650, respectively.

The comparative PSNR results of the CFOPS-BSBEA method is reported in Table4; Fig.7. The results show that the CFOPS-BSBEA model reaches better performance under all images. With SI of IMG 1, the CFOPS-BSBEA technique provides a higher PSNR of 57.35dB whereas the BESOPSCE, CEGAN, GSOSM, SSOSM, and CSOSM models attain lower PSNR of 55.32dB, 51.70dB, 51.32dB, 50.81dB, and 49.93dB, respectively. Also, With SI of IMG 2, the CFOPS-BSBEA model delivers a greater PSNR of 58.11dB whereas the BESOPSCE, CEGAN, GSOSM, SSOSM, and CSOSM models attain lower PSNR of 55.95dB, 52.93dB, 51.99dB, 50.42dB, and 50.24dB, respectively.

The comparative SSIM results of the CFOPS-BSBEA method are reported in Table5; Fig.8. The results show that the CFOPS-BSBEA model gets better performance under all images. With SI of IMG 1, the CFOPS-BSBEA technique attains a higher SSIM of 0.9996 whereas the BESOPSCE, CEGAN, GSOSM, SSOSM, and CSOSM models attain lower SSIM of 0.9992, 0.9934, 0.9782, 0.9684, and 0.9565, respectively. Also, With SI of IMG 2, the CFOPS-BSBEA model delivers a greater SSIM of 0.9990 whereas the BESOPSCE, CEGAN, GSOSM, SSOSM, and CSOSM models attain lower SSIM of 0.9983, 0.9918, 0.9807, 0.9678, and 0.9523, respectively.

Hence, the CFOPS-BSBEA technique can be exploited to improve security in the cloud environment.

Authors and Affiliations

1. Department of Instrumentation Engineering, Madras Institute of Technology Campus, Anna University, Chromepet, Chennai 44, India

   Vikas K. Soman&V. Natarajan

Contributions

Conceptualization, VS; methodology, VS; software, VS; validation, NV; formal analysis, VS; investigation, NV; resources, NV; data curation, NV; writing—original draft preparation, VS; writing—review and editing, VS and NV; visualization, VS and NV; supervision, VS and NV; project administration, NV. All authors have read and agreed to the published version of the manuscript.

Corresponding author

Correspondence to Vikas K. Soman.

Competing interests

The authors declare no competing interests.

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions