Efficient SCAN and Chaotic Map Encryption System for Securing E-Healthcare Images

Abstract

The largest source of information in healthcare during the present epidemic is radiological imaging, which is also one of the most difficult sources to interpret. Clinicians today are forced to rely heavily on therapeutic image analysis that has been filtered and sometimes performed by worn-out radiologists. Transmission of these medical data increases in frequency due to patient overflow, and protecting confidentiality, along with integrity and availability, emerges as one of the most crucial components of security. Medical images generally contain sensitive information about patients and are therefore vulnerable to various security threats during transmission over public networks. These images must be protected before being transmitted over this network to the public. In this paper, an efficient SCAN and chaotic-map-based image encryption model is proposed. This paper describes pixel value and pixel position manipulation based on SCAN and chaotic theory. The SCAN method involves translating an image’s pixel value to a different pixel value and rearranging pixels in a predetermined order. A chaotic map is used to shift the positions of the pixels within the block. Decryption follows the reverse process of encryption. The effectiveness of the suggested strategy is evaluated by computing the histogram chi-square test, MSE, PSNR, NPCR, UACI, SSIM, and UQI. The efficiency of the suggested strategy is demonstrated by comparison analysis. The results of analysis and testing show that the proposed program can achieve the concept of partial encryption. In addition, simulation experiments demonstrate that our approach has both a faster encryption speed and higher security when compared to existing techniques.

1. Introduction

Digital medical images are becoming more and more important in today’s hospitals in diagnosing and treating diseases as a result of the pandemic, and, consequently, they are garnering more attention. Medical photos are stored and transmitted across networks while also containing the patient’s diagnostic data. With the development of technology, it is now much quicker and simpler to transmit medical information in a pandemic emergency. Unexpected theft, unwanted viewing, or improper use of these private photos could result in catastrophic consequences. The unapproved images could be used for personal advantage, such as medical marketing and false insurance claims, by a hacker or rogue database administrator, for instance, resulting in a potentially fatal outcome. Protecting medical images is essential as a result. It is essential to have a reliable, secure system for transmitting these digital images. Digitization greatly facilitates communication between the doctor and their patients. Doctors from far away can work together and cooperate. E-health has a wide range of applications, such as drug synergy prediction, disease diagnosis, digital surgery, and telemedicine. Health data are commonly transmitted over public networks, which can lead to various security threats.

The transmission of medical images is typically protected through cryptographic algorithms. Image encryption and timely diagnosis ought to take as little time as possible. However, due to limited battery lives and limited processing options, devices with limited resources cannot perform complex encryption processes. As a result, encryption is outsourced to the cloud or another third party. Encryption techniques can be used to safeguard medical images to address these security concerns. Algorithms for encryption, such as the Data Encryption Standard (DES) and Advanced Encryption Standard (AES), are utilized for scrambling text information and are not appropriate for picture encryption [1,2,3,4,5]. Manjula and Mohan [6] chose an important part of a medical image to show the hidden data, which is typically done by selecting the more frequently used parts of the image. An algorithm for using FF1 and FF3-1 to partially encrypt confidential data in images was proposed by Jang and Lee [7]. Without increasing the size of the data, which could result in a waste of storage space, the confidential data will be encrypted. Sankaradass and others [8] present the grayscale encryption method in light of the return on initial capital investment with confusion. Initially, the Sobel edge detection method must be utilized to identify the ROI component. The image must then be divided into important and unimportant sections using the blocks’ edges. The ROI portion is encrypted using the Lorenz system, and the unimportant region is encrypted using Sine maps.

The following are telemedicine’s most pressing issues. The setting and conventional cryptosystems are as follows: (i) medical image confidentiality without sacrificing quality; (ii) the reliability of confidential medical image data; (iii) effective searching without decoding the surrounding medical images distributed; (iv) differential attacks; (v) the robustness of the private key.

It is still a research task to create an efficient cryptographic system with dependent encryption. The translation domain is the foundation of the current selective cryptographic encryption strategy, which is not tightly coupled with conversion coding techniques as with the cryptographic process, which is designed for the conversion of inputs and/or outputs. The majority of them eventually yield low encryption performance. Consequently, this work is expected to provide incomplete encryption innovation that works more successfully, more realistically, securely, and rapidly for constant applications.

The Efficient SCAN and Chaotic Map Encryption System for Secure and Rapid E-Healthcare Image Encryption was proposed in light of all the aforementioned drawbacks of the existing algorithms. This algorithm was created to defeat differential, chosen image, brute force, and plain image attacks. Cryptography is used to verify the algorithm’s effectiveness and robustness. The proposed algorithm and the most recent algorithms are also compared.

Three elements are included in the proposed work to ensure the quick and safe transmission of medical images.

• Medical image that protects patient privacy;
• The integrity of a medical image’s region of interest (tumor, injured area, or MRI impression);
• Secure image retrieval for diagnosis.

The following are the major contributions of the proposed paper:

• To encrypt e-healthcare images, a practical model of image encryption is proposed;
• An efficient permutation–diffusion model based on various SCAN methods is adopted;
• The SCAN method entails converting the pixel value of a picture to a different pixel value and rearranging the pixels in the image in a specific order;
• The pixels in the block are repositioned using a chaotic map;
• The proposed system is able to effectively encrypt the pixels within a particular bounding box by identifying the pixel;
• Combining the advantages of SCAN maps and chaotic maps, a partial encryption system is proposed; the encryption system is sensitive to the initial values and easy to implement.

The remainder of the paper is organized as follows. Section 2 gives the related work. Section 3 explains the various methods involved in the proposed encryption work. Section 4 proposes the effective partial image encryption system architecture for medical images. Section 5 reports the security analysis. Section 6 presents the simulation results and discussion. Section 7 gives conclusions for this work.

2. Related Works

Saeed Bahrami [9] offers Media Content Encoding as a partial DCT transform coefficient encoding scheme. In the frequency transform domain and in the form of partial coding, which is time-efficient, it has a high execution speed and is compatible with the compression structure. Lightweight and quick algorithms are frequently employed to encode multimedia content. Based on three fundamental principles and in opposition to each other, a fast and lightweight proposed method and two designs are introduced in the partial encryption scheme of the DCT coefficient of variation. The partial encoding proposed by Xinjun Zhang et al. [10] is based on the chaos of the image SPIHT color code. The recovered image is attenuated to diverse degrees when the binary values of various styles obtained by the Color-SPIHT (CSPIHT) compression algorithm are modified. This implies that different six-bit types contribute differently to decoding. We can ensure that no information will leak and that very little processing will be required as long as the most important bits that contribute the most are encrypted. Due to its perfect fallacy and confusion, the Piecewise Linear Chaos Map (PWLCM) is a good choice for producing a series of random numbers utilized in the suggested cryptosystem. A partial encoding strategy for media content was proposed by Shiguo Lian [11]. Wavelet transform is used to implement the partial encoding technique. It is secure against current assaults and successful in deployment in the actual world. A straightforward selective coding method based on a saw tooth space fill curve, pixels of interest, a nonlinear chaos map, and single value decay was presented by Gaurav Bhatnagar [12]. This algorithm’s main goal is to blur the positions of the pixels with a skewed space-fill curve before choosing the crucial pixels with the pixels of interest approach. Then, utilizing a secret picture key derived from a nonlinear chaos map and single value decay, diffusion is carried out on the crucial pixels. Based on JPEG2000, Zahia Brahimi et al. [13] presented novel selective encoding image schemes. Only codes belonging to specific sensitive locations are blocked by the first number. We are introducing the swapping of contribution code blocks in a few places to increase security. This approach adheres to the JPEG2000 code stream’s format and does not add extra JPEG2000 markers to the protected code stream. In the wavelet domain, Nidhi et al. [14] have suggested a selective coding method for conditional access systems. To conserve computing time and computational resources, this encoding is selectively applied to a portion of the media data streams rather than the complete media data, managing the multimedia data’s transparency at the time of encryption. A notion was described by Priyanka Agrawal and Manisha Rajpoot [15] where a major piece of an image can be generated by choosing a component of the image, which is then used in the normal operation mode, commonly used for encryption. Encrypted data are delivered along with the rest of the message’s original body after encryption is complete, guaranteeing secure distribution over open networks. The primary goal of the current effort is to select a region of an image by arranging the bit stream into a grid and choosing the grid’s diagonal. Xiang Tao et al. [16] stated that the majority of selective picture encoding techniques are codec-specific because they are built on top of image compression methods. This work proposes a selective grayscale picture encoding strategy because the various bit planes of the image contribute differentially to the viewing effect. Only a portion of each pixel’s significant bits, which are encoded, using the keystream created by one-way concatenating the mapped network, demonstrate good chaotic dynamics in this approach. The selective encoding of the compressed image was described by Marc Van Droogenbroeck and Raphal Benedett [17] using JPEG compression. The Huffman encoder synthesizes the coefficients of zeros into a succession of zeros and employs termination symbols. It matches magnitude types for zero with the string zero. The Huffman encoder assigns these symbols 8-bit code words. The sign and magnitude of the non-zero coefficients are described by the code words that come before the appended bits. The extra bits that correspond to a particular number of AC coefficients are encoded in the suggested technique. It is thought that the DC coefficient contains significant information that is both observable and highly predictable; hence, it is not encoded. According to Roman P Farrhofer and Andreas Uhl [18], the idea of a grayscale image is divided into 8-bit planes, with the most important bit planes being encoded. After some testing, it was discovered that (1) 2-bit plane selective encoding is adequate if the image quality deviates from the data, and (2) 4-bit most significant plane encoding is not secure enough. Seriousness is acceptable, (3) 4-bit plane encryption offers good security, and (4) only the lowest resolution of the five layers can be encrypted for selective encryption. By rearranging the mapped images under SCAN models and choosing a pixel value from the rearranged mapping images under the mapping function, we have presented a combinatorial strategy. An innovative technique that achieves lossless compression and the encryption of binary and grayscale images was presented by S.S. Maniccam and N.G. Bourbakis [19]. The SCAN models produced by the SCAN method serve as the foundation for the compression and encryption schemes. SCAN is a formal language-based technology for two-dimensional spatial access that effectively specifies and generates a collection of spatial pathways or scans to fill curves. An image encryption and decryption algorithm based on the SCAN method was proposed by Chao-Shen Chen and Rongjain Chen [20]. SCAN is a two-dimensional, language-based spatial assessment technique that effectively specifies and generates a collection of scanlines. In this case, the original image is filled with the scan path text to create an encoded image. Image encryption based on chaos Arnold, Logistic, and Hennon maps, among other categories of chaotic cards, depends on the rearranging of image pixels under non-linear properties [21]. To concurrently improve invisibility and robustness, a secure watermarking approach and selective encryption techniques are proposed. For data embedding, the suitable coefficient is determined using the transform domain scheme [22]. Arnold’s Cat Map and the 2D Logistic-Sine-Coupling Map (2D-LSCM) are two chaotic maps used in the chaotic-based image encryption system that increase the randomness and security of the encrypted image [23]. Shima Ramesh Maniyath et al. [24] explained that the goal of the proposed system is to present a sophisticated framework, where a deep neural network has been used to optimize the performance of straightforward encryption techniques. For improved security performance, the chaotic map notion is further incorporated into the robustness of the optimization approach. Lang Li et al. [25] created a novel encryption technique that encrypts using both the SP network structure and the Feistel network structure. The current SP network has a problem because the encryption and decryption methods are distinct. We modified the SP network structure using involution-related features of the linear and nonlinear components to address this issue. The altered one makes it possible for the circuit or program for encryption and decryption to function as the Feistel network structure. Yasir Naseer et al. [26] explained the use of a 3D mixed chaotic map in an innovative method of image encryption. In the suggested scheme, we used a 3D mixed chaotic map to transform the positions of the pixels by permuting them and then mixing them in a row- and column-wise manner. Tengfei Wu et al. [27] evaluated the influence of grayscale images and suggested a model for the FPGA and embedded systems. In the shortest amount of time feasible, a quicker and more accurate disease diagnosis may be achieved with the use of this combination. The study improved image processing by developing cancer diagnosis models that are more precise and effective. Narima Zermi et al. [28] describe a novel strategy that entails exactly meticulously integrating patient information and hospital signature information into the medical image. Our ambitious goal is to accurately incorporate the watermark with minimal distortion so that the medical information in the image is normally retained. This adaptable technique applies DWT decomposition to the image, allowing for an incredibly satisfying modification during insertion. Med Sayah Moad et al. [29] suggest a method of patient identification and watermark integrity checking by watermarking. In this method, the patient’s encrypted imagery is included in the second section of the watermark, which also carries the patient’s information fingerprint. The medical image is split into four sub-bands using a discrete wavelet transform for integration. The resultant mid-frequency coefficients are then modulated to integrate the watermark bits. The capability of this method permits the integration of the patient’s photograph and fingerprint, and it will also be suitable for the potential addition of error-correcting code. The integration process’s coefficient modification allowed for the unnoticeable concealing of the watermark.

Kiran et al. [30] explained block cipher-based region of interest medical picture encryption with various maps, which is suggested in this research paper. First, region of interest (ROI) regions are recovered using a Laplacian edge detection operator, and the picture is then divided into wanted (ROI) and unwanted (RONI or ROB) components depending on the edges in the blocks. Arnold’s cat map and angle value are then used to permute the significant ROI regions circularly. Unimportant regions remain unchanged, while the permuted ROI section is encrypted using the Duffing mechanism. The ROI encrypted component and the region of background (ROB) unmodified part are combined to produce the ROI encrypted image. Zhenlong Man et al. [31] developed an image encryption method by creating random numbers using the Least Squares Generative Adversarial Network (LSGAN). We build a novel learning random number generator using LSGAN’s potent learning capabilities. To achieve the quick and effective production of random numbers, six chaotic systems with various architectures and dimensions are employed as training sets. Aiman Jan et al. [32] introduce an Image Encryption Framework (IEFHAC) that uses the Hessenberg transform and chaotic encryption to encrypt patient data while enhancing security and speeding up computation. IEFHAC uses two 1D chaotic maps, the Sine map and the logistic map, to confuse the data, while the Hessenberg household transform is used to accomplish dispersion. Although the presented solutions offer higher security than the conventional encryption techniques, the algorithm’s execution takes longer. Therefore, the creation of less time-consuming, safe, and trustworthy chaos-based algorithms is required. In Noura M et al.’s work [33], the ROI of medical imaging relies on encryption assuming a replacement operation, encrypting the entire medical image using a permutation operation. Another bitplane-based scheme was developed by Muthu J.S. and others [34]. Four-bit planes with high information content for medical images were selected, sorted, and replaced the entire medical image. Shafique A et al. proposed a three-level structure for encoding medical images [35]. In the first and third levels, four-level information-rich bit planes were chosen so that they can be sorted within bit planes, respectively. Aditya Kumar Sahu et al. [36] proposed an approach that takes into account two adjacent host image (HI) pixels for embedding the watermark bits. Each HI pixel experiences a maximum of one alteration at the embedding end to create the watermarked pixels. In addition, the second technique makes minimal use of distortion by reproducing the picture and the watermark bits using the idea of mirrored images of the HI. Abdul RehmanJaved et al. [37] provided a survey of future technology analysis and smart city requirements. We conduct a thorough study to identify and examine the most recent technological developments, which will form the basis of the next prosperous period. Chandramohan Dhasarathan et al. [38] discuss the homomorphic standard system functionality, which includes all functional requirements for deep learning systems in COVID-19 health management.

3. Methods

In this study, SCAN and chaotic maps used for encrypting medical images efficiently. The SCAN method is used for pixel diffusion and chaotic maps for pixel permutation. Each method is explained in the following section.

3.1. SCAN Methodology

With the help of the SCAN language [20], this technique transforms 2D images into a 1D list. Some SCAN letters are used in this language. One type of scan order is represented by each SCAN letter. Secret images can be created with various SCAN letter combinations. The technique will then produce a SCAN string after determining the SCAN character combination. The original image’s digitization order is specified by this sequence. After scanning the original image in the designated order, this method further encrypts the SCAN string using industry-standard cryptographic tools. The original image is secure, since an unauthorized user cannot obtain the correct SCAN string. Figure 1 shows the different types of SCAN patterns (0–7 represents different types of SCAN patterns).

3.2. Chaotic Map

Map selection is a crucial stage in any chaotic digital encoding. Different behaviors of chaos maps, such as sensitivity to the initial conditions, responsiveness to noise orbital disturbances, chaos intervals, periodic windows, etc., have an impact on the structure or functionality of the chaotic encoding system. It is desirable to offer some independence between the cryptographic system and the map, because some systems have been broken due to failure to take into account the strengths and efficiency of the chosen chaotic map. Due to its independence, a good cryptosystem can satisfy the security and efficiency requirements without needing to fully understand the chosen chaos map [21,22].

The logistic map and the tent map are the two choices due to their mathematical simplicity. The logistic map is as follows:

$$X\_\left(n+1\right)=r\ast X\_n\ast \left(1-X\_n\right)$$

(1)

The logistic map chaotic signal used has primary values of X0 ∈ [0,1] and r ∈ [3.57,4].

4. Proposed Image Encryption Scheme

Figure 2 depicts the block diagram for encoding a portion of an image using random and block concatenation maps. The initial block size is 4 × 4 and the input image is mxn in size. In order to create a partially encoded image, the input image is split into numerous 4 × 4 blocks, and the pixel values of these blocks are mixed together using chaotic maps. The previously partially encrypted image serves as the input image each time, and we select a different block size from the block size list table given in Figure 2 to obtain various partially encrypted images. The following is a description of the encryption process.

Steps for encrypting the medical images are explained below. Partially encrypted images are obtained for different block sizes as shown in

Table 1. Block size list for permutation.

Sl. No.	Block Size
1	4 × 4
2	8 × 8
3	16 × 16
.
.
.
N	(m/2) × (n/2)

.

Step 1: Choose a base mapping image or mapping image based on the SCAN template.

Step 2: Translate each pixel of the input image’s encryption to its corresponding 8-bit binary number.

Step 3: Rearrange the 8-bit binary integer into a variety of upper 4-bit and lower 4-bit components.

Step 4: Convert these two 4-bit bits to their equivalents in decimal notation.

Step 5: Retrieve the gray pixel from the mapped image using these two decimal values. In this case, the picture mapping uses the higher nibble decimal equivalent as the row indicator and the lower nibble decimal equivalent as the column indicator.

Step 6: In order to create a visible encoded image, replace the pixel value of the input image with the pixel value acquired in Step 5 of the image mapping to obtain a visible encoded image.

5. Parameters for the Evaluation

The following parameters are used for the evaluation the performance of the proposed image encryption technique.

5.1. Histogram Analysis Using Chi Square Test

The scattering of pixels in an image’s histogram is shown by their frequency of occurrence. A cypher picture that has a consistent distribution of pixels is more secure. The chi-square (2) test is used to determine whether encrypted images based on mathematical formulas are uniform. The formula for the chi-square test is provided by

$$\chi 2={\displaystyle \sum }_{L=0}^{255}\frac{{\left(observed value-expected value\right)}^2}{expected value}$$

(2)

L represents the degree of gray. Greater regularity of the pixel distribution in the encrypted image is indicated by lower values of 2. In

Table 2. Chi-square test for histogram uniformity analysis.

PIE	1	2	3	4	5	6	7
C SCAN	100.23	120.45	133.12	165.23	201.3036	221.1126	241.0036
D SCAN	100.56	120.67	133.45	165.94	204.5374	224.3585	254.4695
O SCAN	100.98	120.33	133.78	165.39	206.4651	226.4867	226.3957
Z SCAN	100.67	120.54	133.61	165.57	204.5751	224.3757	264.2757
S SCAN	100.12	120.52	133.24	165.89	207.4651	227.4567	247.3767

the chi-square values for the encrypted images are listed.

5.2. Mean Squared Error

Mean squared error (MSE) is the average of the square of the difference between an encrypted image and a plain image [39]. The equation yields the MSE (3).

$$MSE=\frac{1}{MXN}{\displaystyle \sum }_{i=1}^M{\displaystyle \sum }_{j=1}^N{\left[X\left(i,j\right)-Y\left(i,j\right)\right]}^2$$

(3)

5.3. Number of Pixels Change Rate (NPCR)

The NPCR [36] is defined as shown below in Equation (4):

$$NPCR=\frac{{\sum }_{i,j}D\left(i,j\right)}{M\times N} \times 100\%$$

(4)

where D is the bipolar array shown in Equation (5), where C1 and C2 are input and output images, respectively.

$$D\left(i,j\right)=\{\begin{array}{c}1, C1\left(i,j\right)\ne C2\left(i,j\right)\\ 0, otherwise \end{array}$$

(5)

5.4. Peak Signal to Noise Ratio (PSNR)

The peak signal to noise ratio is negatively correlated with MSE and is expressed in dB. Equation (6) below gives the equation for PSNR [39].

$$PSNR=10{log}_{10}\frac{255}{MSE}$$

(6)

5.5. Unified Average Changed Intensity (UACI)

This is a measurement of the rate of intensity difference between the plain image and the cypher image [39]. It is written as Equation (7), which is given below.

$$UACI=\frac{1}{N}\left[{\displaystyle \sum }_{i,j}\frac{\left|C1\left(i,j\right)-C2\left(i,j\right)\right|}{255}\right]$$

(7)

5.6. Structural Similarity Index Matrix (SSIM)

This value indicates how similar any two photos are. The range of the SSIM value is 0 to 1 [40]. The SSIM value will be 1 (the ideal value if the two photos are different); otherwise, it will be 0.

$$SSIM\left(x,y\right)=\left[\frac{\begin{array}{cc}\left(2\mu x\mu y+C1\right)& \left(2\sigma xy+C2\right)\end{array}}{\begin{array}{cc}\left(\mu x^2+\mu y^2+C1\right)& \left(\sigma x^2+\sigma y^2+C2\right)\end{array}}\right]$$

(8)

6. Experimental Results

In this experiment, we used a variety of the 512 × 512 pixel images depicted in Figure 3. Graphs in Figure 4, Figure 5, Figure 6, Figure 7 and Figure 8 display the outcomes of the suggested approach (from 2 to 8). According to Figure 4 and Figure 5, there is less coding according to MSE and PSNR when C-SCAN mapping and swapping are combined with a smaller block size (4 × 4), but there is more coding when C-SCAN and swapping are combined with a smaller block size (4 × 4). O-SCAN mapping and swapping are combined with block size. The number of encodings according to the MSE/PSNR is not specifically within our control, but we can alter the values of MSE/PSNR by selecting the appropriate block size and SCAN patterns.

The combination of C-SCAN mapping and permutation with smaller block sizes (4 × 4) results in a lower NPCR than other combinations, which are almost larger than or equal to 99, according to Figure 6. According to Figure 7, the UACI is lower when C-SCAN mapping and permutation are combined with smaller block sizes and variables than the other combinations. For combinations of C-SCAN mappings and swaps with smaller block sizes, SSIM and UQI, as shown in Figure 8 and Figure 9, are summed and altered for different combinations.

As can be seen from

Table 3. Results obtained from proposed method for baby in womb image.

	1	2	3	4	5	6	7
PEI
C SCAN
D SCAN
O SCAN
Z SCAN
S SCAN

, the encrypted images produced by the suggested methods have a consistent pixel distribution and successfully masked all of the hackers’ data.

We compared some key quantitative index values of the proposed algorithm with other semi-selective encryption algorithms, and the results are shown in

Table 4. Comparison of NPCR value with existing state of the art methods.

Literature	NPCR
[32]	99.61
[33]	99.65
[34]	99.62
[39]	99.83
Proposed SSCAN	99.85

. We find that the proposed algorithm outperforms existing procedures overall or at least not inferiorly on key indicators.

7. Conclusions

The application of information and communication technology (ICT) in healthcare leads to smart health systems. Improved disease diagnosis and treatment may result from the use of technology-based healthcare solutions. Overall, the quality of the treatment given to the patients will improve as a result of this. The security of medical images is crucial to the effective delivery of services in smart e-health systems. To reduce the confidentiality concerns associated with the transfer of sensitive medical data in emergencies, image data security has grown to be a key concern. The effective implementation of smart health infrastructures needs to guarantee the security of medical pictures. In this paper, we have considered how to encrypt an image using a mapping technique based on the SCAN method and a chaotic map to swap blocks of pixels. We conclude that a smaller block size for the permutation will result in higher visibility in the encoded image based on the experimental findings. As the swap block sizes increase, so do the number of encryptions. The system is resistant to attacks thanks to the chaotic sequence generating technique, and since one of the keys is random, it is challenging for the anonymous hacker to determine the precise key for decryption.

The effectiveness of the algorithm has been assessed by evaluating the security as well as the amount of time needed to perform encryption and decryption. The security was computed, compared to the algorithm suggested in the current literature, and results were good. The algorithm’s suitability for use in real applications was also demonstrated in the time complexity study.

Additionally, the SCAN concept converts the muddled image with completely distinct pixel values, enhancing system security. The scheme performs better over time, and, in terms of security, and it can prevent numerous attacks, including differential attacks and statistical attacks. In terms of NPCR, UACI, and computational time, it yields values of 99.66–100%, 37.39%, and 0.23 s, respectively. The proposed method outperforms the currently used techniques. The proposed method may therefore be applied in actual medical settings in smart, healthy cities.