Physiological Signal-based Double Chaotic Encryption Apparatus and Method for Protecting Medical Information for E-Healthcare Services

Abstract

A double chaotic encryption device and method for protecting medical information based on biosignals for E-healthcare services is intended to provide a unique wireless communication protocol for medical and healthcare environments, and can provide a double chaotic encryption method that includes the natural randomness of biosignals. The present invention can guarantee the security of the system and resist multiple attacks such as noise, and in the case of plain text images, the encryption time does not exceed 0.1 seconds, thereby having the effect of guaranteeing the real-time requirements of the system.

Description

Technology Field

[0001] The present invention relates to a double chaotic encryption device, and more specifically, to a double chaotic encryption device and method for protecting biosignal-based medical information for E-healthcare services, which can provide a double chaotic encryption method including the natural randomness of biosignals, for providing a unique wireless communication protocol for medical and healthcare environments. Background Technology

[0002] Wireless Body Area Networks (WBANs) have received significant attention in science and industry in recent years as a fundamental technical support for immediate E-healthcare services. They define unique wireless communication protocols for medical and healthcare situations.

[0003] However, concerns regarding WBAN's network and data security became an obstacle to its rapid development.

[0004] Since unauthorized access to and illegal data tampering of WBANs can cause significant problems for users, standards also recommend Elliptic Curve Public Key Cryptography (ECC) and the Advanced Encryption Standard (AES) as security methods. However, following the rapid growth of wearable and implantable medical and long-term health monitoring devices in recent years, an increasing number of researchers and engineers are questioning whether these traditional methods are suitable security measures for WBANs. The main reason is that limited hardware resources and power supplies can barely support the complex computations of ECC and AES. Consequently, designing and implementing lightweight data encryption methods capable of meeting the requirements of low hardware complexity and high data encryption strength has become a hot topic in the WBAN research community. Prior art literature

[0005] Korean Registration No. 10-1083429 The problem to be solved

[0006] To solve such problems, the present invention aims to provide a unique wireless communication protocol for medical and healthcare environments, and to provide a double chaotic encryption device and method for protecting medical information based on biosignals for E-healthcare services, which can provide a double chaotic encryption method including the natural randomness of biosignals. means of solving the problem

[0007] A dual chaos encryption device for biosignal-based medical information protection for E-healthcare services according to the features of the present invention for achieving the above objective is,

[0008] An EEG (Electroencephalogram) eigenvalue extraction unit that extracts EEG (Electroencephalogram) eigenvalues ​​containing deep multi-domain features as initial factors for a chaos-based pseudo-random number generator to ensure the randomness of the initial value;

[0009] A control unit that generates a hybrid pseudo-random number generator based on PLCM (Piecewise Linear Map), a chaos mapping having pseudo-random number characteristics using the extracted EEG eigenvalues, and LTM (Coupled Logistic-Tent Map), a chaos mapping combining Logistic mapping and Tent mapping; and

[0010] It includes a dual chaos encryption processing unit that applies a dual chaos encryption technique based on a random number sequence generated by the above-mentioned pseudo-random number generator and the eigenvalues ​​of the Electroencephalogram (EG) for the Wireless Body Area Network (WBAN).

[0011] A dual chaos encryption method for protecting medical information based on biosignals for E-healthcare services according to the features of the present invention is

[0012] A step of extracting Electroencephalogram (EEG) eigenvalues ​​containing deep multi-domain features as initial factors for a chaos-based pseudo-random number generator to ensure the randomness of the initial values;

[0013] A step of generating a hybrid pseudo-random number generator based on PLCM (Piecewise Linear Map), a chaos mapping having pseudo-random number characteristics using the extracted EEG eigenvalues, and LTM (Coupled Logistic-Tent Map), a chaos mapping combining Logistic mapping and Tent mapping; and

[0014] It includes the step of applying a double chaos cryptography technique based on a sequence of random numbers generated by the above pseudo-random number generator and the eigenvalues ​​of the Electroencephalogram (EEG) for the Wireless Body Area Network (WBAN). Effects of the invention

[0015] With the above-described configuration, the present invention can ensure the security of the system and resist multiple attacks such as noise, and in the case of a plain text image, the encryption time does not exceed 0.1 seconds, thereby ensuring the real-time requirements of the system.

[0016] The present invention has the effect of enabling the construction of a security system capable of resisting various attacks by solving the security problem of initial values ​​in a chaos device using brainwave eigenvalues ​​extracted from a WBAN that includes the randomness of the real world. Brief explanation of the drawing

[0017] FIG. 1 is a diagram showing the overall configuration of a WBAN and E-healthcare service system according to an embodiment of the present invention. FIG. 2 is a diagram showing the configuration of a dual chaos encryption device for protecting medical information based on biosignals for an E-healthcare service according to an embodiment of the present invention. Figure 3 is a diagram showing a PLCM branching diagram according to an embodiment of the present invention. Figure 4 is a diagram showing an LTM branching diagram according to an embodiment of the present invention. FIG. 5 is a diagram showing six channels of EEG data according to an embodiment of the present invention. FIG. 6 is a diagram showing the correlation between time domain features and frequency domain features according to an embodiment of the present invention. FIG. 7 is a diagram illustrating the process of generating a pseudo-random number generator based on chaos mapping according to an embodiment of the present invention. FIG. 8 is a diagram illustrating a double chaos encryption process according to an embodiment of the present invention. Specific details for implementing the invention

[0018] Throughout the specification, when a part is described as "including" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components.

[0019] The present invention provides a dual chaos encryption method that incorporates natural randomness by adding an electroencephalogram (EEG), a biological signal. Initially, to ensure the natural randomness of the initial values, three-dimensional eigenvalues ​​containing deep multi-domain features are extracted as initial elements of a chaos-based pseudo-random number generator. Then, a hybrid pseudo-random number generator is generated based on two chaos maps, the Piecewise Linear Map (PLCM) and the Coupled Logistic-Tent Map (LTM), which possesses the advantages of high randomness and low hardware complexity. Finally, a dual chaos encryption method based on EEG eigenvalues ​​for WBANs is provided.

[0020] FIG. 1 is a diagram showing the overall configuration of a WBAN and an E-healthcare service system according to an embodiment of the present invention, and FIG. 2 is a diagram showing the configuration of a dual chaos encryption device for protecting medical information based on biosignals for an E-healthcare service according to an embodiment of the present invention.

[0021] The present invention includes a Wireless Body Area Network (WBAN) (100) and an E-healthcare service system (110) based on the WBAN.

[0022] A WBAN (Wireless Body Area Network) (100) is a network that can perform remote medical diagnosis and treatment by collecting biosignals from implantable devices operating inside the human body or wearable sensors operating outside the human body and analyzing the body condition based on the collected biosignals.

[0023] The E-healthcare service system (110) may be a concept that encompasses all institutions, devices, and servers that provide E-healthcare services using body sensing information received from multiple WBANs (100).

[0024] The E-healthcare service system (110) is a system that collectively refers to a remote clinic hospital, pharmacy, ambulance, remote medical service, double chaos encryption device, etc. In the present invention, the double chaos encryption device (120) for protecting medical information based on biosignals for E-healthcare services will be described in detail.

[0025] As illustrated in FIG. 2, a dual chaos encryption device (120) for protecting medical information based on biosignals for an E-healthcare service according to an embodiment of the present invention may include a signal receiving unit (121), a control unit (122), a storage unit (123), an EEG unique value calculation unit (124), an autoencoder generation unit (125), a pseudo-random number generator (126), and a dual chaos encryption processing unit (127).

[0026] The signal receiver (121) can receive each individual's EEG (Electroencephalogram) signal from a plurality of WBANs (100). The EEG signal is one of the most common physiological brainwave signals transmitted from a WBAN.

[0027] Brainwaves are classical biological signals that objectively reflect the neural activity of the human cerebral cortex in the form of electricity; however, their nature is abnormal, and they contain immense uncertainty regarding the real world due to greater individual differences and chaos.

[0028] EEG, one of the most common vital signs for medical and healthcare purposes, has been widely studied over the past few decades. Particularly in recent years, with the development of brain-computer interfaces, brainwave-based research is being successfully applied to real-life situations, such as emotion recognition, epilepsy detection, and driving assistance.

[0029] EEG signals are one of the most common biological signals transmitted from a WBAN.

[0030] The control unit (122) can store the EEG signal received from the signal receiving unit (121) in the storage unit (123).

[0031] The EEG eigenvalue calculation unit (124) calculates EEG eigenvalues ​​having time domain and frequency domain features in the WBAN to ensure the randomness of the initial factor of the dual chaos encryption device (120).

[0032] More specifically, the EEG eigenvalue calculation unit (124) extracts the time domain features and frequency domain features of the raw multi-channel EEG signal in the first step.

[0033] Due to the low power consumption and complexity of implementation characteristics of the dual chaos encryption device (120) of the present invention, the lightweight authentication and encryption method based on a chaos map is widely used in low-power systems such as the Internet of Things (IoT) and wearable devices.

[0034] The present invention provides a dual chaotic stream encryption method based on EEG eigenvalues ​​for WBANs.

[0035] The EEG eigenvalue extracted from the WBAN containing the natural randomness of the real world can be calculated as an initial element to solve the security problem of the initial value in the double chaos encryption device (120).

[0036] The cascade of PLCM and LTM effectively improves the randomness of the generated sequence, and the sequence can pass NIST testing.

[0037] The EEG eigenvalue calculation unit (124) calculates EEG eigenvalues ​​having time domain and frequency domain features in the WBAN to ensure the randomness of the initial value of the dual chaos encryption device (120).

[0038] More specifically, the EEG eigenvalue calculation unit (124) extracts time domain and frequency domain features of the raw multi-channel EEG signal. The signal processing method helps to extract more universal representations from the raw multi-channel EEG and can also reduce computational costs.

[0039] Next, unsupervised learning methods and autoencoders are applied to extract deep features in the time domain and frequency domain, respectively.

[0040] This can accommodate noise inputs that help resolve the unavoidable noise effects during the EEG acquisition process and enables automatic feature selection.

[0041] Finally, deep time-domain features and deep frequency-domain features are combined to form multi-domain features, and then an autoencoder is used to mine the relationship between the time domain and the frequency domain to obtain deep multi-domain features. Deep multi-domain features are used to represent EEG signals.

[0042] The control unit (122) collects EEG data collected at different times from a plurality of WBANs (100) through the signal receiving unit (121) and stores it in the storage unit (123).

[0043] The control unit (122) can acquire EEG data by using a sampling frequency of 200 Hz for subsampling in the case of raw EEG data and filtering out other unrelated frequency bands using a filter of 1 to 75 Hz.

[0044] The EEG eigenvalue calculation unit (124) extracts time domain features and frequency domain features from the EEG data stored in the storage unit (123).

[0045] The EEG eigenvalue calculation unit (124) can extract an EEG (Electroencephalogram) eigenvalue containing deep multi-domain features as an initial factor for a chaos-based pseudo-random number generator to ensure randomness of the initial value.

[0046] The present invention encrypts based on multiple chaotic concepts in the key value used for encryption and the data encryption process to maintain confidentiality.

[0047] The EEG eigenvalue calculation unit (124) can represent a piecewise linear chaos map by the following mathematical formula 1 and a combined logistic tent map by the following mathematical formula 2.

[0048] FIG. 3 is a diagram showing a PLCM branching diagram according to an embodiment of the present invention, and FIG. 4 is a diagram showing an LTM branching diagram according to an embodiment of the present invention.

[0049] Piecewise Linear Chaotic Map (PLCM)

[0050] PLCM is a general one-dimensional chaos mapping and is used in dual chaos encryption systems, and can be expressed as Equation 1 below.

[0051]

[0052] As shown in Figure 3, it illustrates the branching structure of piecewise linear chaos with an initial value y(0) and a structural parameter P.

[0053] The value interval of the initial value y(0) is (0, 1), and p is (0, 1 / 2). At this point, the double chaos encryption system enters a chaotic state, and the sequence generated by the iteration acquires pseudo-random characteristics. At this point, the double chaos encryption system has the following characteristics.

[0054] (1) The output sequence moves within the interval (0,1).

[0055] (2) The output trajectory has a delta-type autocorrelation function.

[0056] (3) It has a uniform and invariant probability density function.

[0057] Coupled Logistic-Tent Map (LTM)

[0058] LTM is a new type of one-dimensional chaos mapping that combines Logistic mapping and Tent mapping, and is as shown in Equation 2 below.

[0059] A random number sequence can be generated by utilizing the LTM as a pseudo-random number generator for a double chaos encryption device.

[0060] The branching diagram of the LTM is shown in Fig. 4. The horizontal axis represents the structural parameter, and the vertical axis represents the 0-1 distribution of the sequence obtained when the structural parameter takes this value. As can be seen in Fig. 4, the sequence is evenly distributed in the interval within the selected range of the structural parameter, which indicates that the sequence has good statistical characteristics.

[0061]

[0062] The EEG eigenvalue calculation unit (124) extracts an EEG signal having time domain features and frequency domain features.

[0063] The channels of the raw multi-channel EEG signal are first selected to reduce the computational cost of multi-channel EEG features and the number of parameters in the deep learning-based model, thereby accelerating the model's convergence speed and ultimately improving the model's performance. Subsequently, time-domain and frequency-domain features are obtained from the EEG signal, which can concisely and powerfully represent the EEG in various domains.

[0064] FIG. 5 is a diagram showing six channels of EEG data according to an embodiment of the present invention.

[0065] The EEG eigenvalue calculation unit (124) captures an EEG (Electroencephalogram) from the cerebral cortex as shown in FIG. 5, and selects six channels of EEG data from the captured EEG.

[0066] Using six channels of EEG data can achieve an effect similar to using all brain electrodes for a given task.

[0067] Accordingly, considering the requirements for wearability, feasibility, and real-time performance for actual application, six EEG data channels named FT7, T7, TP7, FT8, T8, and TP8, indicated by circles in Fig. 5, are selected.

[0068] EEG data It is defined as follows. Here, s represents the number of samples, c represents the selected channel, and f represents the sampling point. c is the same as 6.

[0069] The EEG eigenvalue calculation unit (124) extracts time domain features from each EEG data as an EEG representation in the time domain.

[0070] The EEG eigenvalue calculation unit (124) extracts time domain features from each sample of EEG data as an EEG (brainwave) representation in the time domain.

[0071] In the case of, An EEG with time-domain features represented by can be obtained after feature extraction. Here, t represents the length of the time-domain feature vector.

[0072] To improve the computational speed of the deep model and reduce the impact of differences between feature sizes on the deep model, normalization is performed on all features of each EEG data, as shown in the following mathematical equation 3.

[0073]

[0074] Here, and represents the features of the i-th sample before normalization, respectively, and and represents the minimum and maximum values ​​of the features, respectively.

[0075] The EEG eigenvalue calculation unit (124) extracts frequency domain features from each EEG data as an EEG representation in the frequency domain.

[0076] Power Spectral Density (PSD) is widely used for frequency domain features of EEG data and is successfully applied to EEG emotion recognition, Alzheimer's brainwave analysis, sleep detection, and more.

[0077] In the present invention, a short-time fast Fourier transform with non-overlapping 1-second window is applied to calculate the PSD of 5 bands.

[0078] Some important frequency bands are related to brain function. For example, the 1 to 3 Hz band is related to the unconscious, and the 4 to 7 Hz band plays an important role in brain function.

[0079] The role in the unconscious, 8 to 13 Hz is generally associated with a relaxed mental state and is more active in the parietal and occipital lobes, 14 to 30 Hz is associated with an active mental state and is more positive in the frontal cortex, and 30 to 50 Hz is associated with attention, consciousness, and learning processes.

[0080] Accordingly, the EEG eigenvalue calculation unit (124) can calculate the PSD of the above five frequency bands, and as a result, the EEG (brainwave) representation in the frequency domain is is obtained. Here, p represents the value of 5 bands. The EEG (brainwave) representation in the frequency domain is identical to the characteristics in the time domain. The above Equation 3 can be used to normalize the characteristics in the frequency domain.

[0081] The autoencoder generation unit (125) generates a deep autoencoder model for mining the correlation between the time domain and the frequency domain in order to represent the EEG more completely and concisely. The overall framework of this approach is shown in FIG. 6. FIG. 6 is a diagram showing the correlation between time domain features and frequency domain features according to an embodiment of the present invention.

[0082] The autoencoder generation unit (125) uses an autoencoder model (operation as in FIG. 6) to generate time domain features and frequency domain characteristics You can obtain.

[0083] The autoencoder generation unit (125) is and Input this into each autoencoder, and using the fully connected (FC) layer with the minimum number of neurons in the autoencoder, respectively and Outputs deep time domain features and deep frequency domain features representing.

[0084] The autoencoder generation unit (125) uses a fully connected (FC) layer with a minimum number of neurons in the autoencoder, each and Outputs deep time domain features and deep frequency domain features representing.

[0085] The autoencoder generation unit (125) is capable of maintaining the original information of the EEG in the time domain and the frequency domain. and Creates.

[0086] and is in the encoder of the following mathematical formulas 4 and 5 and Creates by assigning.

[0087]

[0088]

[0089] The autoencoder generating unit (125) is for the decoder and If we substitute, as shown in the following mathematical equations 6 and 7, and It can output.

[0090]

[0091]

[0092] Here, represents the operation of the encoder (Encode) and decoder (Decode).

[0093] The autoencoder generation unit (125) uses the following mathematical formula 8 to generate a deep time domain feature and deep frequency domain feature Merging multi-domain features Gets.

[0094]

[0095] Here, silver and Represents multi-domain features obtained by merging.

[0096] The autoencoder generation unit (125) has multi-domain features according to the following mathematical formula 9. input into the encoder to deep multi-domain features printing

[0097]

[0098] Deep multi-domain features used to represent EEG representations utilizing autoencoders to mine correlations and complementarity between multi-domain features Gets.

[0099]

[0100] The autoencoder generation unit (125) can configure a time domain autoencoder, a frequency domain autoencoder, and a multi-domain autoencoder, respectively, because the input dimensions and learning objectives are different.

[0101] The details of the three types of autoencoders are as shown in Table 1 below.

[0102]

[0103] The autoencoder generation unit (125) can obtain multi-domain features of the EEG signal through a time domain autoencoder, a frequency domain autoencoder, and a multi-domain autoencoder.

[0104] The autoencoder is a fully connected (FC) system with the fewest number of neurons.

[0105] In the autoencoder, the activation function is set to Tanh, the loss function is defined as mean squared error, and the optimization tool is Adam.

[0106] Because there are significant differences in EEG between individuals, it is customized for each individual, which helps improve the performance of encryption algorithms.

[0107] Deep multi-domain features It is mapped to the secret key space using mathematical formula 11.

[0108]

[0109] Here, a and b represent the minimum and maximum values ​​in the secret key space, and a and b are 0.1 and 0.5, and and represents the features of the i-th sample before normalization, respectively, and and represents the minimum and maximum values ​​of the features, respectively.

[0110] The control unit (122) generates a pseudo-random number generator (126) based on PLCM (Piecewise Linear Map), a chaos mapping having pseudo-random number characteristics using extracted EEG eigenvalues, and LTM (Coupled Logistic-Tent map), a chaos mapping combining Logistic mapping and Tent mapping.

[0111] FIG. 7 is a diagram illustrating the process of generating a pseudo-random number generator based on chaos mapping according to an embodiment of the present invention.

[0112] The control unit (122) receives an EEG unique value (S100) and can generate a pseudo-random number generator (126), which is a device that generates a pseudo-random number sequence using PLCM (Piecewise Linear Map), a chaos mapping with pseudo-random number characteristics using the EEG unique value, and LTM (Coupled Logistic-Tent Map), a chaos mapping that combines Logistic mapping and Tent mapping (S110, S120, S130).

[0113] In the initial stage of encryption, multi-domain features of the EEG signal can be obtained through three Deep Autoencoders, and the EEG unique value is normalized with the key of the dual encryption device (120) to obtain an initial factor. The EEG signal itself has randomness.

[0114] Using EEG unique values ​​as part of the key can significantly improve the secretivity of the key.

[0115] The present invention provides a dual chaos pseudo-random number generator (126) based on PLCM and LTM, and the number of iterations is the same as that of a single chaos generator. In this algorithm, three EEG eigenvalues ​​are used as keys: the initial value y (0) of PLCM, the structural parameter P, and the structural parameter u of LTM.

[0116] PLCM can obtain a random sequence in the range (0, 1) through iteration.

[0117] The first 20 digits of the PLCM iteration are discarded, and the 21st value becomes the starting valid value. This value is used as the initial value of the LTM, and the number of LTM iterations is set to 10,000. Once the number of LTM iterations is specified, the PLCM iterates once to generate the initial value for the next LTM. Figure 9 shows the flowchart for generating a random sequence. Since the sequence generated by the PLCM is also random, random interference is added to the LTM with the same computational load as a single chaos, thereby guaranteeing the unpredictability of the orbit.

[0118] FIG. 8 is a diagram illustrating a double chaos encryption process according to an embodiment of the present invention.

[0119] The EEG eigenvalue calculation unit (124) extracts an EEG (Electroencephalogram) eigenvalue containing deep multi-domain features as an initial factor for a chaos-based pseudo-random number generator to ensure randomness of the initial value (S200, S210).

[0120] The control unit (122) generates a hybrid pseudo-random number generator based on PLCM (Piecewise Linear Map), a chaos mapping with pseudo-random number characteristics using the extracted EEG eigenvalues, and LTM (Coupled Logistic-Tent Map), a chaos mapping that combines Logistic mapping and Tent mapping (S220).

[0121] The dual chaos encryption processing unit (127) obtains a chaos sequence Xn having randomness in the range (0, 1) by a pseudo-random number generator (126). Then, the binary sequence Sn used for encryption can be determined by a threshold value such as the following mathematical formula 12.

[0122] The dual chaos encryption processing unit (127) obtains a chaos sequence Xn with randomness in the range (0, 1) as the output of the pseudo-random number generator (126), and obtains a binary sequence Sn used for encryption by substituting the chaos sequence Xn into the following mathematical formula 12.

[0123] The double chaos encryption processing unit (127) forms a plaintext represented as I(M, N) into a pixel-unit matrix (M, N) (S230, S240), and converts the pixel-unit matrix (M, N) into a binary sequence In (S250).

[0124] The double chaos encryption processing unit (127) converts the valid information of the plaintext represented by I(M, N) into a binary sequence In with length n = M × N × 8, and obtains a ciphertext sequence stream Zn by performing an XOR operation between a random sequence stream Sn and a plaintext sequence stream In using the following mathematical formula 11 (S260).

[0125] The objective of completely hiding valid information of plaintext represented by I(M, N) from the encryption system can be achieved through the stream encryption technique proposed in the present invention.

[0126] Convert plain text into a binary sequence In of length n = M × N × 8. This is the length of the required chaos sequence.

[0127] The ciphertext sequence stream Zn can be obtained by XORing the random sequence stream Sn and the plaintext sequence stream In using the following mathematical formula 13.

[0128] The ciphertext sequence stream Zn to be transmitted can be obtained by comparing the random sequence stream Sn with the plaintext sequence stream In. Equation 13 illustrates a double chaos-based encryption process.

[0129]

[0130]

[0131] Although embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements by those skilled in the art using the basic concept of the present invention as defined in the following claims also fall within the scope of the present invention. Explanation of the symbols

[0132] 100: WBAN 110: E-Healthcare Service System 120: Double Chaos Encryption Device

Claims

Claim 1 In a dual chaos encryption device, an Electroencephalogram (EEG) eigenvalue extraction unit that extracts Electroencephalogram (EEG) eigenvalues ​​containing deep multi-domain features as initial factors for a chaos-based pseudo-random number generator to ensure randomness of the initial value; and a control unit that generates a hybrid pseudo-random number generator based on PLCM (Piecewise Linear Map), a chaos mapping having pseudo-random number characteristics, and LTM (Coupled Logistic-Tent Map), a chaos mapping combining Logistic mapping and Tent mapping, using the extracted EEG eigenvalues. A dual chaos encryption device for protecting biosignal-based medical information for E-healthcare services, comprising a dual chaos encryption processing unit that applies a dual chaos encryption technique based on a random number sequence generated by the hybrid pseudo-random number generator and an EEG (Electroencephalogram) eigenvalue for a WBAN (wireless body area network); wherein the EEG eigenvalue extraction unit captures an EEG (Electroencephalogram) from the cerebral cortex, selects channels of six EEG data—the EEG data are defined by the following EEG formula 1—from the captured EEG, extracts time domain features from each EEG data as an EEG representation in the time domain—the EEG representation in the time domain is defined by the following EEG formula 2—and extracts frequency domain features from each EEG data as an EEG representation in the frequency domain—the EEG representation in the frequency domain is defined by the following EEG formula 3— Here, s is the number of samples, c is the selected channel, f is the sampling point, and c is 6.[Equation 2] Here, t is the length of the time-domain feature vector.[Equation 3] Here, p is the value of the 5 frequency bands. Claim 2 A dual chaos encryption device for biosignal-based medical information protection for an E-healthcare service, wherein the EEG eigenvalue extraction unit extracts an EEG signal having time domain features and frequency domain features, connects the time domain features and the frequency domain features to form a multi-domain feature, and then uses an autoencoder to mine the relationship between the time domain region and the frequency domain region to obtain a deep multi-domain feature for representing the EEG signal. Claim 3 delete Claim 4 In claim 1, an autoencoder model is generated for mining the correlation between the time domain feature and the frequency domain feature, and the time domain feature is generated using the generated autoencoder model. and the above frequency domain characteristics Obtaining, the above and the above Each is input into an autoencoder, and the autoencoder uses a fully connected (FC) layer with the minimum number of neurons to each and A dual chaos encryption device for protecting biosignal-based medical information for E-healthcare services, further comprising an autoencoder generator that outputs deep time domain features and deep frequency domain features representing [Equation 4]. [Mathematical Formula 5] [Mathematical Formula 6] [Mathematical Formula 7] Here, is an operation of the encoder and decoder. Claim 5 In claim 4, the autoencoder generation unit uses the following mathematical formula 8 to obtain the deep time domain feature and the above deep frequency domain feature Merging multi-domain features Step to obtain; and multi-domain features by the following mathematical formula 9 input into the encoder to deep multi-domain features Dual chaos encryption device for biosignal-based medical information protection for E-healthcare services outputting [Equation 8] Here, silver and It is a multi-domain feature obtained by merging.[Equation 9] Claim 6 A dual chaos encryption device for protecting biosignal-based medical information for an E-healthcare service, wherein the dual chaos encryption processing unit obtains a chaos sequence Xn having randomness in the range (0, 1) as the output of the hybrid pseudo-random number generator, obtains a binary sequence Sn used for encryption by substituting the chaos sequence Xn into the following mathematical formula 10, converts valid information of plaintext represented as I(M, N) into a binary sequence In having a length n = M × N × 8, and obtains a ciphertext sequence stream Zn by performing an XOR operation on the random sequence stream Sn and the plaintext sequence stream In using the following mathematical formula 11. [Mathematical Formula 10] [Mathematical Formula 11] Claim 7 A step of extracting Electroencephalogram (EEG) eigenvalues ​​containing deep multi-domain features as initial factors for a chaos-based pseudo-random number generator to ensure randomness of the initial values; a step of generating a hybrid pseudo-random number generator based on PLCM (Piecewise Linear Map), a chaos mapping having pseudo-random characteristics, and LTM (Coupled Logistic-Tent Map), a chaos mapping combining Logistic mapping and Tent mapping, using the extracted EEG eigenvalues; A dual chaos encryption method for protecting biosignal-based medical information for E-healthcare services, comprising the step of applying a dual chaos encryption technique based on a random number sequence generated by the hybrid pseudo-random number generator and the eigenvalues ​​of an Electroencephalogram (EG) for a Wireless Body Area Network (WBAN), wherein the step of extracting as an initial factor further comprises: a step of capturing an Electroencephalogram (EG) from the cerebral cortex; a step of selecting channels of six EEG data from the captured EEG—where the EEG data is defined by the following Equation 1; a step of extracting time domain features from each EEG data as an EEG representation in the time domain—where the EEG representation in the time domain is defined by the following Equation 2; and a step of extracting frequency domain features from each EEG data as an EEG representation in the frequency domain—where the EEG representation in the frequency domain is defined by the following Equation 3. [Equation 1] Here, s is the number of samples, c is the selected channel, f is the sampling point, and c is 6.[Equation 2] Here, t is the length of the time-domain feature vector.[Equation 3] Here, p is the value of the 5 frequency bands. Claim 8 A dual chaos encryption method for biosignal-based medical information protection for E-healthcare services, wherein the step of extracting the initial factor further comprises: a step of extracting an EEG signal having time domain features and frequency domain features; and a step of forming a multi-domain feature by connecting the time domain feature and the frequency domain feature, and then using an autoencoder to mine the relationship between the time domain region and the frequency domain region to obtain a deep multi-domain feature for representing the EEG signal. Claim 9 delete Claim 10 In claim 7, the step of generating an autoencoder model for mining the correlation between the time domain feature and the frequency domain feature; and the time domain feature using the generated autoencoder model and the above frequency domain characteristics Step of obtaining; and the above and the above Each is input into an autoencoder, and the autoencoder uses a fully connected (FC) layer with the minimum number of neurons to each and A dual chaos encryption method for protecting biosignal-based medical information for E-healthcare services, further comprising a step of outputting deep time domain features and deep frequency domain features representing [Equation 4]. [Mathematical Formula 5] [Mathematical Formula 6] [Mathematical Formula 7] Here, is an operation of the encoder and decoder. Claim 11 In claim 10, the deep time domain feature using the following mathematical formula 8 and the above deep frequency domain feature Merging multi-domain features Steps to obtain; and multi-domain features By inputting into the encoder, deep multi-domain features as shown in Equation 9 below A dual chaos encryption method for biosignal-based medical information protection for E-healthcare services, which further includes a step of outputting.[Equation 8] Here, silver and It is a multi-domain feature obtained by merging.[Equation 9] Claim 12 A dual chaos encryption method for protecting biosignal-based medical information for E-healthcare services, wherein the step of generating the hybrid pseudo-random number generator comprises: generating the autoencoder model as a time domain autoencoder, a frequency domain autoencoder, and a multi-domain autoencoder; and further comprising the step of obtaining multi-domain features of an EEG signal through the time domain autoencoder, the frequency domain autoencoder, and the multi-domain autoencoder. Claim 13 The method for double chaos encryption for biosignal-based medical information protection for E-healthcare services according to claim 7, wherein the step of applying the double chaos encryption technique comprises: obtaining a Chaotic Sequence Xn having randomness in the range (0, 1) as the output of the hybrid pseudo-random number generator; obtaining a Binary Sequence Sn used for encryption by substituting the Chaotic Sequence Xn into the following Equation 10; converting valid information of plaintext represented as I(M, N) into a Binary Sequence In having a length n = M × N × 8; and further comprising the step of obtaining a ciphertext sequence stream Zn by performing an XOR operation on the random sequence stream Sn and the plaintext sequence stream In using the following Equation 11. [Equation 10] [Mathematical Formula 11] Claim 14 In claim 5, the deep multi-domain features A dual chaos encryption device for biosignal-based medical information protection for E-healthcare services, mapped to a secret key space using the following mathematical formula 12.[Mathematical Formula 12] Here, a and b represent the minimum and maximum values ​​in the secret key space, respectively, and and represents the features of the i-th sample before normalization, respectively, and and Each represents the minimum and maximum values ​​of the feature. Claim 15 In claim 11, the deep multi-domain features A double chaos encryption method for protecting biosignal-based medical information for E-healthcare services, mapped to a secret key space using the following mathematical formula 12.[Mathematical Formula 12] Here, a and b represent the minimum and maximum values ​​in the secret key space, respectively, and and represents the features of the i-th sample before normalization, respectively, and and Each represents the minimum and maximum values ​​of the feature.

Images