Medical image compression and encryption method and system based on memristor FHN neuron model
Through the medical image compression encryption method based on the memristor FHN neuron model, the problem of insufficient security of traditional Chinese medicine image encryption in the prior art is solved, and an efficient and secure medical image encryption and decryption process is realized.
Patent Information
- Application Number
- CN202510411873.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-02
- Publication Date
- 2025-06-10
AI Technical Summary
There is a problem of insufficient security in existing medical image encryption schemes based on memristor neuron models.
The medical image compression encryption method based on the memristor FHN neuron model is adopted. By fusing medical images and sparselying discrete wavelet transformation, the initial conditions and control parameters of the m-FHN chaotic system are generated using the SHA-512 hash value, and the chaotic sequence is iteratively generated for diffusion and chaos processing to obtain a compressed encrypted image.
It significantly improves the security of encrypted images, significantly improves the ability to resist differential attacks, strong binding of keys to plain text, reduces compression rate by 50%, reduces storage requirements, and supports multimodal medical image format.
Smart Images

Figure CN120128662A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to, but is not limited to, the field of image encryption technology, and particularly relates to a medical image compression and encryption method and system based on a memristive FHN neuron model. Background Art
[0002] The secure transmission of medical imaging data has become a core challenge in the construction of intelligent healthcare. According to the HIPAA Journal's 2024 annual report, the proportion of medical images involved in global medical data breaches has climbed from 37.6% in 2020 to 61.2% in 2023. Among them, 83% of the breaches stem from unencrypted or weakly encrypted algorithms during the transmission process. Images such as mammograms and PET-CTs contain biometric identifiers. Experiments show that deep learning can be used to recover patient identity information from encrypted images with an accuracy of up to 79.3%. This has prompted new encryption technologies that integrate biological neurodynamics to become a research hotspot.
[0003] Neuronal electrophysiological models provide new ideas for constructing highly secure encryption systems. The classical FitzHugh-Nagumo (FHN) model simplifies the Hodgkin-Huxley equation and abstracts ion channel dynamics into a two-dimensional differential equation system. Although this model can simulate the firing patterns of neuronal action potentials (such as periodic spikes and chaotic oscillations), its dynamic characteristics have fundamental limitations. The maximum Lyapunov exponent λ_max of the two-dimensional system is ≤0.15, resulting in a Kolmogorov entropy of only 1.28 bits / ms for the phase trajectory on the plane. As a result, the proportion of the generated pseudo-random sequences passing the NIST SP800-22 test is less than 82%, making it difficult to meet the strict randomness requirements of medical image encryption.
[0004] The introduction of memristors provides a physical basis for breaking through the dynamic limitations of traditional neuron models. As described by Chua's theory, the flux-charge nonlinear relationship of memristors (φ(q) = αq + βq^3) can accurately simulate the weight adaptation characteristics of neural synapses. The 3D M-FHN system constructed by embedding memristors into the FHN model exhibits the dynamic characteristics of coexisting heterogeneous attractors and can arbitrarily switch between spike bursts, periodic, and chaotic discharge behaviors, with very complex dynamic characteristics, providing a solid theoretical basis for its application in the field of medical image encryption.
[0005] In view of the above analysis, the technical problems that urgently need to be solved in the existing technology are: there are few reported medical image encryption schemes based on memristive neuron models, and there are also problems with insufficient security. Summary of the Invention
[0006] In view of the problems existing in the prior art, the present invention provides a medical image compression and encryption method and system based on a memristive FHN neuron model.
[0007] The present invention is implemented as follows. A medical image compression and encryption method based on a memristive FHN neuron model includes:
[0008] S1, fusing n medical images of any size, whether in color or grayscale, into a fused image of size M×N×3;
[0009] S2, performing sparse representation on the fused medical image in combination with discrete wavelet transform to further reduce redundant data;
[0010] S3, generating the initial conditions and control parameters of the m-FHN chaotic system through the SHA-512 hash value of the original image to ensure strong correlation between the key and the plaintext;
[0011] S4, iterating the m-FHN system to generate three groups of chaotic sequences, and performing quantization processing on the chaotic sequences to obtain a pseudo-random matrix for diffusion and scrambling operations;
[0012] S5, successively performing forward diffusion, index scrambling, and backward diffusion operations on the sparse image to obtain the final compressed and encrypted image;
[0013] S6, decryption operation: The receiving end obtains the initial parameters and control parameters bound to the plaintext through a secure channel, and then successively performs inverse backward diffusion, inverse index scrambling, inverse forward diffusion, and inverse discrete wavelet transform on the ciphertext image to obtain the decrypted image.
[0014] Furthermore, an m-FHN neuron model is constructed. The model is a bistable system, which can exhibit periodic states or chaotic states depending on different initial values, and the complex dynamic characteristics of the model are verified by analyzing the bifurcation diagram and Lyapunov exponent diagram of parameter k1.
[0015] Furthermore, image fusion is performed on the plaintext image, and then the key is substituted into the model according to the fused image to obtain chaotic sequences; then, through quantization processing, pseudo-random matrices X1, X2, and Y1 are obtained to scramble and diffuse the fused image P(M×N×3) to obtain the encrypted image.
[0016] Furthermore, the forward diffusion process is as follows:
[0017]
[0018] Where A represents the image matrix obtained after forward diffusion processing, i and j represent the number of rows and columns of matrix A, mod represents the modulo function, P represents the fused image, and X1 represents the pseudo-random matrix.
[0019] The image matrix A is obtained through forward diffusion processing, and then index scrambling is performed:
[0020] A1(i,:) = A(Y1(i,:),:) (2)
[0021] The image matrix A1 is obtained through index scrambling, and finally the backward diffusion processing is performed to obtain the final compressed and encrypted image C:
[0022]
[0023] Among them, cr represents the compression ratio, and M and N respectively represent the length and width of the fused image.
[0024] Furthermore, the ciphertext image C is successively subjected to inverse backward diffusion, inverse index scrambling, inverse forward diffusion, and inverse discrete wavelet transform to obtain the decrypted image P1.
[0025] Another object of the present invention is to provide a medical image compression and encryption system based on a memristive FHN neuron model, including:
[0026] Receiving module: Receives medical images (supporting formats such as DICOM, JPEG2000, etc.), and after preprocessing, transmits them to the encryption module.
[0027] Pseudo-random generation module: The core is a three-dimensional m-FHN chaotic system, and a highly random key is generated by integrating the SHA-512 hash function.
[0028] Encryption module: Performs sparsification, diffusion, and scrambling processing in sequence.
[0029] Decryption module: Executes the encryption steps in reverse, supporting end-to-end decryption and restoration.
[0030] Output module: Outputs the encrypted image to cloud storage or a transmission channel.
[0031] Combined with the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solutions to be protected by the present invention are:
[0032] (1) Efficiency improvement: Integrated compression and encryption processing greatly saves processing time and is suitable for real-time transmission scenarios. The reconstruction reduces the data volume by 50%, and the storage requirement is reduced.
[0033] (2) Enhanced security: The key is strongly bound to the plaintext, and the ability to resist differential attacks is significantly improved (NPCR > 99.6%, UACI > 33.4%). Index scrambling destroys the pixel spatial correlation (the correlation coefficient approaches 0).
[0034] (3) Compatibility and Practicality: Support multi-modal medical image formats and adapt to medical cloud platforms and edge computing devices. The encrypted image histogram is uniformly distributed and passes the NIST SP 800-22 randomness test. Description of the Drawings
[0035] Figure 1 is a flowchart of a medical image compression and encryption method based on a memristive FHN neuron model provided by an embodiment of the present invention;
[0036] Figure 2 is a coexistence graph of a chaotic attractor and a periodic attractor provided by an embodiment of the present invention;
[0037] Figure 3 is a bifurcation diagram of parameter k1 provided by an embodiment of the present invention;
[0038] Figure 4 is a Lyapunov exponent diagram of parameter k1 provided by an embodiment of the present invention;
[0039] Figure 5 is a schematic diagram of plaintext, ciphertext, and decrypted image provided by an embodiment of the present invention;
[0040] Figure 6 is a histogram of plaintext and ciphertext images provided by an embodiment of the present invention;
[0041] Figure 7 is a correlation diagram of plaintext and ciphertext images provided by an embodiment of the present invention;
[0042] Figure 8 is to apply a small perturbation of 10 to parameter k1 in an embodiment of the present invention -15 to obtain a schematic diagram of a decryption failure image;
[0043] Figure 9 is a structural diagram of a medical image compression and encryption system based on a memristive FHN neuron model provided by an embodiment of the present invention. Detailed Embodiments
[0044] In order to make the objectives, technical solutions, and advantages of the present invention clearer, the following further elaborates on the present invention in conjunction with embodiments. It should be understood that the specific embodiments described herein are merely used to explain the present invention and are not used to limit the present invention.
[0045] Aiming at the limitations of existing medical image security technologies, the present invention aims to propose a new medical image compression and encryption method based on a memristive FHN neuron model to systematically solve the technical defects of traditional solutions in terms of anti-attack performance and processing efficiency.
[0046] (1) First aspect: Image encryption and decryption method
[0047] 1) Fuse n pairs of medical images of any size, whether in color or grayscale, into a fused image of size M×N×3.
[0048] 2) Combine discrete wavelet transform to sparsely represent the fused medical image, further reducing redundant data.
[0049] 3) Generate the initial conditions and control parameters of the m-FHN chaotic system through the SHA-512 hash value of the original image to ensure a strong correlation between the key and the plaintext.
[0050] 4) Iterate the m-FHN system to generate three groups of chaotic sequences, and perform quantization processing on the chaotic sequences to obtain pseudo-random matrices for diffusion and scrambling operations.
[0051] 5) Perform forward diffusion, index scrambling, and backward diffusion operations on the sparsified image in sequence to obtain the final compressed and encrypted image.
[0052] 6) Decryption operation steps: The receiving end obtains the initial parameters and control parameters bound to the plaintext through a secure channel, and then performs inverse backward diffusion, inverse index scrambling, inverse forward diffusion, and inverse discrete wavelet transform on the ciphertext image in sequence to obtain the decrypted image.
[0053] Step 1: Construct an m-FHN neuron model. The model is a bistable system and can exhibit periodic or chaotic states depending on the initial values, as Figure 2 shown. The complex dynamic characteristics of the model are verified by analyzing the bifurcation diagram and Lyapunov exponent diagram of parameter k1, as Figure 3 and Figure 4 shown.
[0054] Step 2: Perform image fusion on the plaintext image, and then substitute the key provided by the fused image into the model to obtain chaotic sequences; then, through quantization processing, obtain pseudo-random matrices X1, X2, Y1 to scramble and diffuse the fused image P(M×N×3) to obtain the encrypted image.
[0055] Among them, the forward diffusion process is as follows:
[0056]
[0057] After forward diffusion processing, obtain the image matrix A, and then perform index scrambling:
[0058] A1(i,:) = A(Y1(i,:),:) (2)
[0059] After index scrambling, obtain the image matrix A1, and finally perform backward diffusion processing to obtain the final compressed and encrypted image C:
[0060]
[0061] Among them, cr represents the compression ratio, and M and N represent the length and width of the fused image respectively.
[0062] Step 3: Perform inverse backward diffusion, inverse index scrambling, inverse forward diffusion, and inverse discrete wavelet transform on the ciphertext image C in sequence to obtain the decrypted image P1.
[0063] Step 4: Analyze the effect of the encryption method through different parameter indicators
[0064] (1) Encryption and decryption effect
[0065] Figure 5 The simulation results of applying the compression encryption method of the present invention to 4 medical images are shown. From top to bottom, they are the original image, the encrypted image, and the image after decryption and reconstruction. The encryption and hiding effects of the image are verified through simulation. It can be seen that the encrypted image has completely encrypted the original image, and the image after decryption and reconstruction can basically restore the original image, indicating the effectiveness of the encryption and decryption method in the present invention.
[0066] (2) Histogram analysis
[0067] Figure 6 This is the histogram corresponding to the original image and the encrypted image in Embodiment 1; it can be seen from the histogram of the original image that it reflects that the distribution of the overall pixel values is very uneven; while the pixel values of the encrypted image are evenly distributed and lose the original statistical characteristics, which indicates that the encrypted image is difficult to be cracked.
[0068] (3) Correlation analysis
[0069] Figure 7 This is the correlation analysis of the original image and the encrypted image. Specifically, the correlation of pixel pairs in the horizontal direction, vertical direction, and diagonal direction is analyzed. It can be seen from the figure that the adjacent pixels of the original image have a high correlation, while the adjacent pixels of the encrypted image have a low correlation; the correlation value of the adjacent pixels of the encrypted image is very close to 0, which indicates that there is almost no correlation between the adjacent pixels of the encrypted image.
[0070] (4) Key sensitivity analysis
[0071] Figure 8 This is the decrypted image obtained after applying a small perturbation to the parameter k1. It can be seen that when the key is slightly changed, the obtained decrypted image shows a noise-like pattern and no effective information about the original image can be distinguished, indicating that the encryption and decryption algorithm provided by the present invention has extremely strong key sensitivity and can effectively resist brute-force attacks. -15 This is the decrypted image obtained after applying a small perturbation to the parameter k1. It can be seen that when the key is slightly changed, the obtained decrypted image shows a noise-like pattern and no effective information about the original image can be distinguished, indicating that the encryption and decryption algorithm provided by the present invention has extremely strong key sensitivity and can effectively resist brute-force attacks.
[0072] (2) As Figure 9 shown, the second aspect: Image encryption and decryption system
[0073] 1) Receiver module: Receives medical images (supporting formats such as DICOM, JPEG2000, etc.), and after preprocessing, transmits them to the encryption module.
[0074] 2) Pseudo-random generation module: The core is a three-dimensional m-FHN chaotic system, which generates a highly random key by integrating the SHA-512 hash function.
[0075] 3) Encryption module: Performs sparsification, diffusion, and scrambling processes in sequence.
[0076] 4) Decryption module: Executes the encryption steps in reverse, supporting end-to-end decryption and restoration.
[0077] 5) Output module: Outputs the encrypted image to cloud storage or a transmission channel.
[0078] (3) Third aspect: Computer-readable medium and electronic device
[0079] 1) Computer-readable medium: Stores encryption and decryption programs, and when executed, calls the m-FHN chaotic system and the compressive sensing algorithm to achieve full-process processing.
[0080] 2) Electronic device: Integrates this system and is applicable to medical imaging devices (such as CT, MRI), remote medical terminals, and mobile devices.
[0081] The image encryption and decryption system first receives standardized medical image data through the receiver module, supporting various imaging formats including DICOM (Digital Imaging and Communications in Medicine), JPEG2000, etc. The received image data undergoes a preprocessing process, including denoising, grayscale normalization, and size standardization, to ensure the consistency and controllability of the data structure during subsequent encryption. The preprocessed image matrix is transformed into a unified numerical representation form that can be input into the encryption algorithm.
[0082] This system uses a three-dimensional chaotic system based on an improved FitzHugh–Nagumo (m-FHN) neuron model as a pseudo-random number generator. Its dynamic behavior exhibits initial value sensitivity and ergodicity. By integrating the SHA-512 hash function to extract the initial feature information of the image to generate a seed value, it further drives the m-FHN system to generate a highly random and unpredictable pseudo-random sequence. This sequence serves as the core source for key generation, ensuring the irreversibility and security of the encryption process.
[0083] The image encryption stage is mainly divided into three steps: sparsification, diffusion, and scrambling. Sparsification uses discrete wavelet transform (DWT) to extract the main structural features of the image, achieving data dimensionality reduction and information hiding; diffusion processing combines forward diffusion and backward diffusion to disrupt the gray-scale distribution and resist statistical attacks; the scrambling operation uses the generated chaotic key sequence to randomly rearrange the spatial position information of the image pixels, significantly enhancing the encryption strength and anti-attack ability of the image.
[0084] The decryption module sequentially performs the inverse scrambling operation, inverse diffusion processing, and sparse reduction and reconstruction in the reverse order of the encryption steps. The decryption process requires the use of the exact same m-FHN initial key and SHA-512 seed value as in the encryption stage to ensure that the decrypted image is highly consistent with the original image. The system supports an end-to-end decryption process, recovering the ciphertext image from the cloud or local device and restoring it to the standard image format to ensure the confidentiality and availability of medical data during remote transmission or storage.
[0085] The encrypted image is transmitted to the specified storage path or secure transmission channel (such as VPN, SSL, or encrypted UDP protocol) through the output module, supporting docking with the hospital's PACS system and remote medical cloud platform. The module is equipped with an image integrity verification and transmission error detection mechanism to enable secure and reliable interaction of the image between different terminal devices. The output format is flexible, supporting image stream encapsulation (such as DICOM encapsulation) to ensure compatibility in medical scenarios.
[0086] The algorithms and logical structures of this system are encapsulated in a computer-readable medium (such as flash memory, SSD, or embedded chip), which can automatically call the m-FHN chaotic model and compressive sensing module during execution to achieve full-process image encryption and decryption operations. The deployed electronic devices include medical imaging devices (CT, MRI), remote consultation terminals, and mobile diagnosis and treatment devices, which have edge computing capabilities and image processing units and can directly complete local encryption or remote decryption of images to meet the application requirements of high-security medical data transmission.
[0087] Example 1
[0088] Please refer to Figure 1 , Figure 1 which is the brief flowchart of the medical image compression and encryption method based on the memristive FHN neuron model in Embodiment 1 of the present invention. The specific steps are as follows:
[0089] Step 1: Construct the m-FHN neuron model. The model is a bistable system and can exhibit a periodic state or a chaotic state depending on the initial value, as Figure 2 shown. The complex dynamic characteristics of the model are verified by analyzing the bifurcation diagram and Lyapunov exponent diagram of parameter k1, as Figure 3 and Figure 4 shown.
[0090] Step 2: Perform image fusion on the plaintext image, then substitute the key provided by the fused image into the model to obtain a chaotic sequence; then, through quantization processing, obtain pseudo-random matrices X1, X2, and Y1, and perform scrambling and diffusion on the fused image P(M×N×3) to obtain the encrypted image.
[0091] Among them, the forward diffusion process is as follows:
[0092]
[0093] After forward diffusion processing, obtain the image matrix A, and then perform index scrambling:
[0094] A1(i,:) = A(Y1(i,:),:) (2)
[0095] After index scrambling, obtain the image matrix A1, and finally perform backward diffusion processing to obtain the final compressed encrypted image C:
[0096]
[0097] Among them, cr represents the compression ratio, and M and N respectively represent the length and width of the fused image.
[0098] Step 3: Perform inverse backward diffusion, inverse index scrambling, inverse forward diffusion, and inverse discrete wavelet transform on the ciphertext image C in sequence to obtain the decrypted image P1.
[0099] Step 4: Analyze the effect of the encryption method through different parameter indicators
[0100] (5) Encryption and decryption effects
[0101] Figure 5 The simulation results of applying the compression encryption method of the present invention to 4 medical images are shown. From top to bottom, they are the original image, the encrypted image, and the image after decryption and reconstruction. The encryption and hiding effects of the image are verified through simulation. It can be seen that the encrypted image has completely encrypted the original image, and the image after decryption and reconstruction can basically restore the original image, indicating the effectiveness of the encryption and decryption method in the present invention.
[0102] (6) Histogram analysis
[0103] Figure 6 The histograms corresponding to the original image and the encrypted image in Example 1 are shown. It can be seen from the histogram of the original image that it reflects that the distribution of the overall pixel values is very uneven; while the pixel values of the encrypted image are evenly distributed, losing the original statistical characteristics, indicating that the encrypted image is difficult to crack.
[0104] (7) Correlation analysis
[0105] Figure 7 For the correlation analysis of the original image and the encrypted image, specifically analyze the correlation of pixel pairs in the horizontal, vertical, and diagonal directions. It can be seen from the figure that adjacent pixels in the original image have a high correlation, while adjacent pixels in the encrypted image have a low correlation; the correlation value of adjacent pixels in the encrypted image is very close to 0, indicating that there is almost no correlation between adjacent pixels in the encrypted image.
[0106] (8) Key sensitivity analysis
[0107] Figure 8 For the decrypted image obtained after applying a small perturbation to the parameter k1 by 10 -15 It can be seen that when the key is slightly changed, the obtained decrypted image shows a noise-like pattern and no effective information about the original image can be distinguished, indicating that the encryption and decryption algorithm provided by the present invention has extremely strong key sensitivity and can effectively resist brute-force attacks.
[0108] The present invention constructs a memristive FHN neuron model with bistability, and uses it to generate high-complexity chaotic sequences to compress and encrypt medical images. At the same time, different chaotic sequences are used in the encryption step, thus generating higher security. This encryption algorithm has high sensitivity to the key, a large key space, can effectively resist various external attacks, and has good practical application prospects.
[0109] I. The specific application fields or related products of the present invention.
[0110] Please refer to Figure 1 , Figure 1 which is the brief flowchart of the medical image compression and encryption method based on the memristive FHN neuron model in Embodiment 1 of the present invention. The specific steps are as follows:
[0111] Step 1: Construct an m-FHN neuron model. The model is a bistable system and can exhibit a periodic state or a chaotic state according to different initial values, as Figure 2 shown. And the complex dynamic characteristics of the model are verified by analyzing the bifurcation diagram and Lyapunov exponent diagram of the parameter k1, as Figure 3 and Figure 4 shown.
[0112] Step 2: Perform image fusion on the plaintext image, and then substitute the key provided by the fused image into the model to obtain a chaotic sequence; then, after quantization processing, obtain pseudo-random matrices X1, X2, Y1 and perform scrambling and diffusion on the fused image P(M×N×3) to obtain an encrypted image.
[0113] Among them, the forward diffusion process is:
[0114]
[0115] The image matrix A is obtained through forward diffusion processing, and then index scrambling is performed:
[0116] A1(i,:) = A(Y1(i,:),:) (2)
[0117] The image matrix A1 is obtained through index scrambling, and finally the backward diffusion processing is performed to obtain the final compressed and encrypted image C:
[0118]
[0119] Among them, cr represents the compression ratio, and M and N represent the length and width of the fused image respectively.
[0120] Step 3: Perform inverse backward diffusion, inverse index scrambling, inverse forward diffusion, and inverse discrete wavelet transform on the ciphertext image C in sequence to obtain the decrypted image P1.
[0121] II. Evidence related to the technical effects obtained in the embodiments of the present invention.
[0122] 1) Encryption and decryption effects
[0123] Figure 5 The simulation results of applying the compression and encryption method of the present invention to 4 medical images are shown. From top to bottom, they are the original image, the encrypted image, and the image after decryption and reconstruction. The encryption and hiding effects of the image are verified through simulation. It can be seen that the encrypted image has completely encrypted the original image, and the image after decryption and reconstruction can basically restore the original image, indicating the effectiveness of the encryption and decryption method in the present invention.
[0124] 2) Histogram analysis
[0125] Figure 6 The histograms corresponding to the original image and the encrypted image in this Embodiment 1 are shown. It can be seen from the histogram of the original image that the distribution of the overall pixel values is very uneven; while the pixel values of the encrypted image are evenly distributed, losing the original statistical characteristics, which indicates that the encrypted image is difficult to be cracked.
[0126] 3) Correlation analysis
[0127] Figure 7 The correlation analysis of the original image and the encrypted image is shown. Specifically, the correlations of pixel pairs in the horizontal direction, vertical direction, and diagonal direction are analyzed. It can be seen from the figure that the adjacent pixels of the original image have a high correlation, while the adjacent pixels of the encrypted image have a low correlation; the correlation value of the adjacent pixels of the encrypted image is very close to 0, which indicates that there is almost no correlation between the adjacent pixels of the encrypted image.
[0128] 4) Key sensitivity analysis
[0129] Figure 8 After applying a small perturbation to parameter k1 by 10 -15 For the decrypted image obtained after the small perturbation, it can be seen that when the key changes slightly, the decrypted image shows a noise-like pattern and no effective information about the original image can be distinguished. This indicates that the encryption and decryption algorithm provided by the present invention has extremely strong key sensitivity and can effectively resist brute-force attacks.
[0130] The present invention constructs a memristive FHN neuron model with bistability and uses the high-complexity chaotic sequence generated by it to compress and encrypt medical images. At the same time, different chaotic sequences are used in the encryption step, thus generating a high level of security. This encryption algorithm has high sensitivity to keys, a large key space, can effectively resist various external attacks, and has good practical application prospects.
[0131] Application Example 1: Secure Transmission of Remote Medical Images
[0132] In a remote medical system, medical images (such as MRI, CT, X-ray films) need to be transmitted between hospitals, clinics, and remote medical platforms. However, traditional encryption methods are vulnerable to attacks and the large amount of data affects the transmission efficiency. The present invention first fuses multiple images into a fused image of M×N×3 and uses discrete wavelet transform for sparsification to reduce redundant data. Subsequently, a chaotic key for the memristive FHN neuron model is generated based on the SHA-512 hash value, and encryption and compression are completed through forward diffusion, index scrambling, and backward diffusion, making the images both secure and efficient. The images only exist in ciphertext form during storage and transmission. Even if the data is intercepted, the original content cannot be restored through statistical analysis or brute-force cracking.
[0133] At the remote medical platform end, doctors can obtain the correct key through a secure channel and sequentially perform inverse diffusion, inverse scrambling, and inverse wavelet transform to accurately restore the medical images, ensuring that the image quality is not damaged. This method significantly enhances the security of medical images, avoids unauthorized access, reduces the bandwidth requirements for remote transmission, reduces the hospital cloud storage cost, and improves the real-time performance of image transmission and remote diagnosis.
[0134] Application Example 2: Secure Protection of Medical Image Cloud Storage
[0135] As medical institutions widely adopt cloud storage systems to manage massive medical image data, data security issues have become a key challenge. Traditional storage methods usually rely on server-side encryption. However, once the server is attacked, there is still a risk of image data leakage. In the present invention, before the images are uploaded to the cloud storage, a unique key is generated through the SHA-512 hash function, and a highly randomized encryption process is achieved in combination with the FHN chaotic system, so that each image has a unique security protection mechanism. The images are encrypted when stored, and unauthorized users cannot directly access the original images. Even if the server is compromised, the data is still unreadable, ensuring the privacy and security of patients.
[0136] When doctors or researchers need to access the medical images stored in the cloud, they can obtain the decryption key through the authorization mechanism and sequentially perform inverse diffusion, inverse scrambling, and inverse transformation to restore clear medical images for diagnosis. This method not only improves the storage security of medical images but also reduces the storage space occupancy through wavelet transform and data compression, reduces the hospital's cloud computing resource costs, and at the same time improves the efficiency of image retrieval and remote access, achieving a dual improvement in security and storage optimization.
[0137] The above has shown and described the basic principles, main features, and advantages of the present invention. For those skilled in the art, it is obvious that the present invention is not limited to the details of the above exemplary embodiments, and without departing from the spirit or basic features of the present invention, the present invention can be implemented in other specific forms. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting. The scope of the present invention is defined by the appended claims rather than the above description. Therefore, all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be encompassed within the present invention. Any reference signs in the claims should not be construed as limiting the claims involved.
[0138] It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated designed hardware. Those of ordinary skill in the art can understand that the above devices and methods can be implemented using computer-executable instructions and / or included in processor control code, such as provided on a carrier medium such as a disk, CD, or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The devices and their modules of the present invention can be implemented by hardware circuits of programmable hardware devices such as very large scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or field programmable gate arrays, programmable logic devices, etc., can also be implemented by software executed by various types of processors, or can be implemented by a combination of the above hardware circuits and software such as firmware.
[0139] As described above, the above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention, any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention should be covered within the protection scope of the present invention.
Claims
1. A medical image compression encryption method based on a memristor FHN neuron model, characterized in that: include: 1) Image fusion: Merge multiple medical images into a fused image of size M×N×3 to improve data processing efficiency; 2) Image sparsification: Discrete wavelet transform (DWT) is used to decompose the fused medical image into low-frequency and high-frequency sub-bands to reduce data redundancy; 3) Chaotic sequence generation: Based on the SHA-512 hash value of the original image, the initial conditions and control parameters of the m-FHN neuron model are calculated, and three sets of chaotic sequences are iteratively generated; 4) Encryption processing: The chaotic sequence is used for forward diffusion, so that the pixel value changes dynamically during the encryption process; Perform index scrambling to rearrange image pixels according to specific rules and break the original spatial association; Backward diffusion is performed to make the image finally form an encrypted state and further reduce the predictability of information; 5) Decryption process: The receiving end obtains the key through a secure channel, and performs inverse backward diffusion, inverse index scrambling, inverse forward diffusion, and inverse discrete wavelet transform in sequence to restore the medical image.
2. The method according to claim 1, characterized in that The m-FHN neuron model is a bistable system, whose neuron state is affected by parameter changes and presents a periodic or chaotic state under different initial conditions. The model exhibits complex dynamic characteristics within different parameter ranges through nonlinear feedback regulation of external input signals and internal variables, ensuring the high security and unpredictability of the encryption system.
3. The method according to claim 1, characterized in that The diffusion and scrambling operations specifically include: 1) Forward diffusion: Using the first set of chaotic sequences, according to the set diffusion rules, each pixel value is affected by the adjacent pixels, thereby enhancing the data obfuscation; 2) Index scrambling: Based on the second set of chaotic sequences, the image pixel positions are rearranged according to the randomly generated index matrix, so that the spatial structure of the image data is completely disrupted; 3) Backward diffusion: The third set of chaotic sequences is used to diffuse the index-scrambled image again, so that the image data can further enhance its anti-attack capability in an encrypted state.
4. The method according to claim 1, characterized in that: The key generation method uses the SHA-512 hash function to calculate the hash value of the plaintext image and maps the hash value to the initial conditions and control parameters of the chaotic system. The specific process is as follows: 1) Extract the pixel information of the plaintext image and calculate the SHA-512 hash value; 2) The first part of the hash value is used to initialize the initial state of the neurons of the chaotic system; 3) The latter part of the hash value is used to set the control parameters of the chaotic system so that the key has a high correlation with the image content; 4) Generate a dynamic key sequence through a chaotic system to ensure that the same encryption input always corresponds to a unique encryption result, preventing known plaintext attacks.
5. The method according to claim 1, characterized in that The security of the encryption method described is enhanced by: 1) Histogram analysis: The pixel values of the encrypted image are evenly distributed, with no obvious identifiable statistical features; 2) Correlation analysis: The correlation between adjacent pixels of the encrypted image is close to zero, which avoids information leakage; 3) Key sensitivity test: When the key changes slightly (for example, the control parameter changes by a very small value), the decrypted image presents a completely unrecognizable noise pattern, ensuring the uniqueness and security of the key; 4) Anti-attack capability: The encryption system can effectively resist known plaintext attacks, chosen ciphertext attacks and statistical analysis attacks.
6. The method according to claim 1, characterized in that The decryption process of the medical image specifically includes: 1) Key acquisition: The receiving end receives the initial key and control parameters through a secure channel; 2) Reverse diffusion: Inverse back diffusion: Use the key to restore the pixel value and eliminate the randomness introduced by the diffusion process; Reverse index scrambling: Arrange pixels in the reverse order of the scrambled index matrix to restore the image to its original arrangement; Reverse forward diffusion: Use chaotic sequence to perform decryption diffusion and restore pixel values to their original state; 3) Inverse discrete wavelet transform: The processed image data is reconstructed through wavelet to restore the original medical image.
7. A medical image compression and encryption system based on a memristor FHN neuron model for implementing the medical image compression and encryption method based on a memristor FHN neuron model as claimed in any one of claims 1 to 6, characterized in that: include: Receiving module: receives medical images and transmits them to the encryption module after preprocessing; Pseudo-random generation module: The core is a three-dimensional m-FHN chaotic system, which generates highly random keys by integrating the SHA-512 hash function; Encryption module: performs sparseness, diffusion and scrambling processing in sequence; Decryption module: reverses the encryption steps and supports end-to-end decryption and restoration; Output module: outputs encrypted images to cloud storage or transmission channels.
8. A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the medical image compression and encryption method based on the memristor FHN neuron model as described in any one of claims 1 to 6.
9. A computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to execute the steps of the medical image compression and encryption method based on the memristor FHN neuron model as claimed in any one of claims 1 to 6.
10. An information data processing terminal, comprising the medical image compression and encryption system based on the memristor FHN neuron model as claimed in claim 7.
Citation Information
Cited By
Breast lesion area identification system and method based on big data
CN121213474A
Anti-interference perception encryption method for unmanned aerial vehicle image
CN122293803A
