CN-121985134-A - Image encryption method and image decryption method based on heterogeneous neural network synchronization with uncertainty

CN121985134ACN 121985134 ACN121985134 ACN 121985134ACN-121985134-A

Abstract

The invention provides an image encryption method and an image decryption method based on heterogeneous neural network synchronization with uncertainty, and relates to the technical field of image processing. The method comprises the steps of obtaining at least two sets of chaotic sequences generated by a driving system in a synchronous state under the condition that the driving system and a response system are synchronous, fusing the at least two sets of chaotic sequences to generate a composite chaotic sequence, and encrypting an original image based on the composite chaotic sequence to generate an encrypted image. At least two sets of response chaotic sequences generated by the response system in a synchronous state are obtained, at least two sets of chaotic response sequences are fused to generate a composite response sequence, and then the encrypted image is decrypted by utilizing the encrypted reverse process. According to the embodiment of the invention, the composite key with more complex dynamic characteristics and stronger randomness is generated by fusing a plurality of groups of chaotic sequences generated by the heterogeneous neural network with uncertainty in a synchronous state, so that the security intensity of an image encryption system is improved.

Inventors

ZHOU LIQUN
GUO YUXIAN
ZHAO ZHIXUE
LIU NA

Assignees

天津师范大学

Dates

Publication Date: 20260505
Application Date: 20260109

Claims (12)

1. An image encryption method based on heterogeneous neural network synchronization with uncertainty, comprising: under the condition that the driving system and the response system are synchronous, at least two groups of chaotic sequences generated by the driving system in a synchronous state are obtained; Fusing the at least two sets of chaotic sequences to generate a composite chaotic sequence; encrypting the original image based on the composite chaotic sequence to generate an encrypted image; The driving system is a first neural network with proportional time lag and uncertainty parameters, the response system is a second neural network without time lag with uncertainty parameters, the internal structures of the first neural network and the second neural network are different, and a heterogeneous neural network with uncertainty is formed through the first neural network and the second neural network.
2. The method of image encryption based on heterogeneous neural network synchronization with uncertainty of claim 1, wherein prior to the step of fusing the at least two sets of chaotic sequences to generate a composite chaotic sequence, the method comprises: applying an adaptive controller on the response system, and driving the state of the response system to realize synchronization with the state of the driving system to obtain a heterogeneous neural network synchronization result; The self-adaptive controller at least comprises a global polynomial synchronous controller, a global exponential synchronous controller and a global asymptotic synchronous controller, and the heterogeneous neural network synchronous result at least comprises a global polynomial synchronous result, a global exponential synchronous result and a global asymptotic synchronous result.
3. The method for encrypting the image based on the synchronization of the heterogeneous neural network with uncertainty according to claim 2, wherein the fusing the at least two sets of chaotic sequences to generate the composite chaotic sequence comprises: acquiring a first chaotic sequence of the driving system based on the global polynomial synchronization result; acquiring a second chaotic sequence of the driving system based on the global index synchronization result; And fusing the first chaotic sequence and the second chaotic sequence to generate the composite chaotic sequence for encryption.
4. The method for encrypting an image based on heterogeneous neural network synchronization with uncertainty according to claim 2, wherein the encrypting an original image based on the complex chaotic sequence to generate an encrypted image comprises: decomposing the original image into a plurality of initial color channel matrices; Disturbing the image pixel positions of the original image based on the composite chaotic sequence and the plurality of initial color channel matrixes to obtain a disturbed color channel matrix; and encrypting the image pixel value of the original image based on the composite chaotic sequence and the disturbed color channel matrix to obtain the encrypted image.
5. The method for encrypting an image based on heterogeneous neural network synchronization with uncertainty according to claim 4, wherein said scrambling the image pixel positions of the original image based on the complex chaotic sequence and the plurality of initial color channel matrices to obtain a scrambled color channel matrix comprises: Designing a nonlinear driving sequence generation algorithm based on the composite chaotic sequence, and generating a pseudo-random sequence according to the nonlinear driving sequence generation algorithm; and respectively carrying out pixel position rearrangement processing on each initial color channel matrix through the pseudo-random sequence to obtain a disturbed color channel matrix.
6. The method for encrypting an image based on heterogeneous neural network synchronization with uncertainty according to claim 5, wherein encrypting the image pixel values of the original image based on the complex chaotic sequence and the scrambled color channel matrix to obtain the encrypted image comprises: Generating a random sequence, and carrying out disturbance processing on the composite chaotic sequence by utilizing the random sequence to obtain a chaotic disturbance driving sequence; intercepting a stable part of the chaotic disturbance driving sequence according to a preset position parameter to obtain a target chaotic disturbance driving sequence; Generating an encryption key matrix corresponding to each color channel based on the target chaotic disturbance driving sequence; performing bit exclusive OR operation on the scrambled color channel matrix and the encryption key matrix of the corresponding color channel to obtain a target encryption channel matrix corresponding to each color channel, and performing combination processing based on the target encryption channel matrix corresponding to each color channel to generate the encrypted image.
7. An image decryption method based on heterogeneous neural network synchronization with uncertainty applied to the implementation of the image encryption method based on heterogeneous neural network synchronization with uncertainty as claimed in any one of claims 1 to 6, characterized by comprising: Under the condition that the same self-adaptive controller as the encryption is used for realizing the synchronization of a driving system and a response system, at least two sets of response chaotic sequences generated by the response system when the synchronization is realized are obtained; Fusing the at least two sets of response chaotic sequences to generate a composite response chaotic sequence; Based on the composite response chaotic sequence, carrying out decryption processing on the received encrypted image to obtain a decrypted image; The driving system is a first neural network with proportional time lag and uncertainty parameters, the response system is a second neural network without time lag with uncertainty parameters, the internal structures of the first neural network and the second neural network are different, and a heterogeneous neural network with uncertainty is formed through the first neural network and the second neural network.
8. The method of image decryption based on heterogeneous neural network synchronization with uncertainty of claim 7, wherein before the step of fusing the at least two sets of response chaotic sequences to generate a composite response chaotic sequence, the method comprises: applying an adaptive controller on the response system, and driving the state of the response system to realize synchronization with the state of the driving system to obtain a heterogeneous neural network synchronization result; The self-adaptive controller at least comprises a global polynomial synchronous controller, a global exponential synchronous controller and a global asymptotic synchronous controller, and the heterogeneous neural network synchronous result at least comprises a global polynomial synchronous result, a global exponential synchronous result and a global asymptotic synchronous result.
9. The method for decrypting an image based on heterogeneous neural network synchronization with uncertainty of claim 8, wherein the fusing the at least two sets of response chaotic sequences to generate a composite response chaotic sequence comprises: acquiring a first response chaotic sequence of the response system based on the global polynomial synchronization result; acquiring a second response chaotic sequence of the response system based on the global index synchronization result; and fusing the first response chaotic sequence and the second response chaotic sequence to generate the composite response chaotic sequence for decryption.
10. The method for decrypting the image based on the synchronization of the heterogeneous neural network with uncertainty according to claim 9, wherein the decrypting the received encrypted image based on the composite response chaotic sequence to obtain the decrypted image comprises: Carrying out channel decomposition processing on the received encrypted image to obtain a plurality of target encrypted channel matrixes; intercepting a stable part of the composite response chaotic sequence according to position parameters used during encryption to obtain a target composite response chaotic sequence; Generating a pseudo-random sequence through a nonlinear driving sequence generation algorithm and the target composite response chaotic sequence; Based on the pseudo-random sequence, performing inverse pixel position rearrangement processing on the multiple target encryption channel matrixes to obtain a color channel matrix with the pixel positions restored; And performing decryption processing based on the color channel matrix after recovering the pixel position and the composite response chaotic sequence to obtain a decrypted image.
11. The method for decrypting an image based on heterogeneous neural network synchronization with uncertainty as claimed in claim 10, wherein said generating a pseudo random sequence by a nonlinear drive sequence generation algorithm and the target complex response chaotic sequence comprises: Designing a nonlinear driving sequence generation algorithm based on the composite response chaotic sequence; and processing the target composite response chaotic sequence through the nonlinear driving sequence generation algorithm to generate a pseudo-random sequence identical to that in encryption.
12. The method for decrypting the image based on the synchronization of the heterogeneous neural network with uncertainty according to claim 11, wherein the decrypting the image based on the color channel matrix after recovering the pixel positions and the composite response chaotic sequence comprises: the random sequence which is the same as the random sequence in encryption is utilized to carry out disturbance processing on the composite response chaotic sequence, and a disturbed composite response chaotic sequence is obtained; generating a decryption key matrix corresponding to each color channel based on the disturbed composite response chaotic sequence; Performing bit exclusive OR operation on the decryption key matrix corresponding to each color channel and the color channel matrix after recovering the pixel position, recovering to obtain a plurality of initial color channel matrixes, and performing combination processing on the initial color channel matrixes to obtain a decrypted image.

Description

Image encryption method and image decryption method based on heterogeneous neural network synchronization with uncertainty Technical Field The invention relates to the technical field of image processing, in particular to an image encryption method and an image decryption method based on heterogeneous neural network synchronization with uncertainty. Background Neural network synchronization techniques are capable of producing chaotic sequences with high complexity and unpredictability, which makes them an ideal source for building dynamic keys in the field of image encryption. The current mainstream method is to establish a driving-responding system based on a homogeneous neural network with the same structure, and encrypt an image by using a generated chaotic sequence after the system is synchronous. The method introduces dynamic characteristics into the encryption process, and improves the security of the traditional static encryption scheme. However, the prior art solutions have two main drawbacks at the source of key generation. First, most studies are based on the ideal case of a "homogeneous" scene with the same model structure of the drive and response system, and the certainty of the parameters. In reality, there are often inherent differences in manufacturing tolerances, environmental disturbances or modeling errors in the system, and uncertainty and structural heterogeneity of the system are ignored by existing schemes, limiting the applicability of the system in actual complex environments. Second, existing methods typically extract a single chaotic sequence from only one synchronization state (e.g., asymptotic synchronization) to generate a key in the construction of the key material. Such key generation mechanisms are relatively simple, the complexity and randomness of the generated keys are limited, and their potential security strength may be inadequate in the face of increasingly advanced cryptanalysis approaches. Therefore, how to utilize uncertainty of system parameters and how to break through theoretical limitation of a homogeneous system, effectively utilize neural networks with different structures and with uncertainty to realize synchronization, and on the basis, design a method capable of fusing more dimensions and richer dynamics characteristics of the system to generate a high-strength composite key becomes a key technical problem for improving safety, robustness and practicability of an image encryption scheme based on the neural networks. Disclosure of Invention The invention provides an image encryption method and an image decryption method based on the synchronization of an uncertainty heterogeneous neural network, which are used for generating a composite key with more complex dynamic characteristics and stronger randomness by fusing a plurality of groups of chaotic sequences generated by the uncertainty heterogeneous neural network in a synchronous state, so that the safety intensity of an image encryption system is fundamentally improved. In a first aspect, the present invention provides an image encryption method based on heterogeneous neural network synchronization with uncertainty, including the steps of: under the condition that the driving system and the response system are synchronous, at least two groups of chaotic sequences generated by the driving system in a synchronous state are obtained; Fusing the at least two sets of chaotic sequences to generate a composite chaotic sequence; encrypting the original image based on the composite chaotic sequence to generate an encrypted image; The driving system is a first neural network with proportional time lag and uncertainty parameters, the response system is a second neural network without time lag with uncertainty parameters, the internal structures of the first neural network and the second neural network are different, and a heterogeneous neural network with uncertainty is formed through the first neural network and the second neural network. In a second aspect, the present invention further provides an image decryption method based on heterogeneous neural network synchronization with uncertainty, applied to the image encryption method based on heterogeneous neural network synchronization with uncertainty in the first aspect, including: Under the condition that the same self-adaptive controller as the encryption is used for realizing the synchronization of a driving system and a response system, at least two sets of response chaotic sequences generated by the response system when the synchronization is realized are obtained; Fusing the at least two sets of response chaotic sequences to generate a composite response chaotic sequence; Based on the composite response chaotic sequence, carrying out decryption processing on the received encrypted image to obtain a decrypted image; The driving system is a first neural network with proportional time lag and uncertainty parameters, the response system is a second neural network