Drug addict identification method and device and storage medium
Through differential entropy extraction of EEG signal characteristics and combining incremental width learning, the existing drug addiction recognition method is solved, and the problem of high cost and single drug type is achieved, efficient identification and real-time monitoring of multiple drugs is achieved, and suitable for large-scale drugs. Scale screening and edge computing deployment.
Patent Information
- Application Number
- CN202410956905.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-17
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2044-07-17
AI Technical Summary
The existing drug addictive identification methods are costly, single drug types are used, complex data preprocessing, and the inability to effectively monitor new drugs, resulting in a decrease in model accuracy and difficulty in deploying in actual large-scale testing.
Differential entropy extracts dynamic features of EEG signals, combined with incremental width learning, and trains through differential entropy features input into incremental width learning model for rapid adaptation and efficient training, avoiding the consumption of medium and high computing resources in deep learning, and is suitable for edge computing deploy.
It improves the accuracy and real-time identification of drug addictives, can effectively monitor a variety of drug types, is suitable for large-scale screening, reduces computing resource consumption, and realizes edge computing deployment.
Smart Images

Figure CN118975804A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of drug addict identification, and in particular to a drug addict identification method, device and storage medium. Background Art
[0002] In the field of drug control and drug rehabilitation law enforcement, the methods for evaluating drug use or addiction mainly rely on behavioral observation, questionnaires, and biological sample testing. Although behavioral observation and questionnaires can obtain self-reports and behavioral performance of subjects to a certain extent, the subjective attitudes and memory biases of the parties involved often affect the objectivity and accuracy of the evaluation results. Biological sample testing can detect biomarkers of drug use behavior, but it takes a lot of time and economic costs, and it is easy for drug users to use deceptive means to evade detection. Therefore, it is necessary to use advanced scientific and technological means and data-driven methods, such as EEG signal analysis, to achieve accurate, objective and dynamic monitoring of drug use behavior and addiction treatment effects.
[0003] Compared with other current EEG-based identification methods, they are too expensive, target a single type of drug, and have complex data preprocessing, making them difficult to use in practice. In addition, the practice of directly inputting EEG signals into the model causes the model to learn other information and noise in the EEG signals, ignoring the characteristics of dynamic changes in EEG signals, resulting in low classification accuracy. In addition, the wet electrode EEG caps used in most methods are not portable, are complicated to wear, and have high equipment costs, making them difficult to deploy in actual testing, which in turn increases the workload.
[0004] At present, with the emergence of new types of drugs, most model samples trained based on EEG data have fewer types of drugs and are unable to learn the characteristics of new drugs. Therefore, in actual large-scale screening, as the number of drug types and sample size increases, the model accuracy will decrease. To solve this problem, it is necessary to re-collect samples and train the model. The training time is long and it is difficult to deploy in actual large-scale testing. Summary of the invention
[0005] In order to solve at least one of the technical problems existing in the prior art to a certain extent, the purpose of the present invention is to provide a method, device and storage medium for identifying drug addicts.
[0006] The first technical solution adopted by the present invention is:
[0007] A method for identifying a drug addict comprises the following steps:
[0008] Acquiring EEG data; the EEG data includes EEG data of healthy persons and EEG data of drug addicts;
[0009] Preprocessing the EEG data to obtain EEG signals;
[0010] Extracting dynamic features of the EEG signal using differential entropy to obtain EEG signal features;
[0011] The obtained EEG signal features are input into incremental width learning for training, and the results are verified by ten-fold cross validation.
[0012] Further, the preprocessing includes at least one of data filtering processing, data segmentation processing, baseline correction processing, or artifact and motion artifact removal processing.
[0013] Furthermore, the preprocessing of the EEG data to obtain EEG signals includes:
[0014] The 0.5-45 Hz filtering technology is used to remove noise and unnecessary frequency components in the EEG data, and the filtered EEG data is divided into time periods of preset time lengths.
[0015] Furthermore, the extracting dynamic features of the EEG signal using differential entropy to obtain EEG signal features includes:
[0016] For each time period of the EEG signal, calculate its first-order difference sequence;
[0017] Estimate the probability density function of a first-difference sequence;
[0018] Using the estimated probability density function, calculate the differential entropy;
[0019] After calculating the differential entropy value of each time period, the differential entropy value is used as the EEG signal feature.
[0020] Furthermore, the first-order difference sequence is obtained by calculating the difference between the values of the EEG signals at adjacent time points;
[0021] The probability density function is estimated by using a histogram method or a kernel density estimation method.
[0022] Furthermore, the calculation formula of the differential entropy is as follows:
[0023]
[0024] Where p(x) represents the probability density function of the signal, and x represents the value of the differential signal.
[0025] Furthermore, the step of inputting the obtained EEG signal features into incremental width learning for training includes:
[0026] The learning process of width learning:
[0027] Input feature data X, first undergo random mapping to form mapping feature node Z n ≡[Z 1 ,Z 2 ,…,Z n ]; where the i-th group of mapping features Z i The calculation of is shown in formula (1):
[0028]
[0029] In the formula, is a linear mapping function, W ei is a random weight, β ei is a random bias, n is the number of groups of mapping features;
[0030] The mapping feature node Z obtained by n After the nonlinear mapping function, the enhanced feature node H is formed m ≡[H 1 ,H 2 ,…,H m ]; the jth enhanced feature H j The calculation of is shown in formula (2):
[0031] H j ≡ζ(Z n W hj +β hj ),j=1,2,…,m (2)
[0032] Where W hj , β hj is the connection weight, m is the number of enhanced nodes, ζ(Z,W hj ,β hj ) is a nonlinear activation function;
[0033] The two features are combined to form a new addiction EEG recognition feature A = [Z|H], and connected to the model output layer Y;
[0034] Since it is known whether each sample has taken drugs, it is only necessary to calculate the connection weight W between the feature layer and the output layer. The calculation formula is as follows:
[0035] W=A -1 Y (3)
[0036] During calculation, ridge regression is used to solve the weight matrix, that is, W is solved by the optimization problem of formula (4):
[0037]
[0038] Among them, σ 1 =σ 2 =v=u=2, the solution is:
[0039] W=(λI+AA T ) -1 A T Y (5)
[0040] In the formula, Y is the true output, I is the unit matrix, λ is the optimization coefficient, and T represents the transpose;
[0041] The width learning prediction output is:
[0042]
[0043] The learning process of data increment width learning:
[0044] For the newly added input data {X a ,Y a}, and calculate the corresponding incremental mapping features and enhanced nodes according to formulas (7) and (8):
[0045]
[0046]
[0047] In the formula, Represents a mapping feature node, represents an enhanced feature node; is the mapping feature node after increment;
[0048] Then the new feature layer is:
[0049]
[0050] In the formula, It is the non-incremental feature layer;
[0051] Output layer weights Updated by the following formula:
[0052]
[0053] in, is the label corresponding to the data Ba; T represents transposition; represents the unincremented feature layer; represents the unincremented output layer weights.
[0054] Furthermore, the calculation formula for the confidence level corresponding to the width learning prediction output Y is as follows:
[0055]
[0056] In the formula, represents the output of the sparse width living body recognition network, Represents the output corresponding to the jth category determined by the sparse width living body recognition network.
[0057] The second technical solution adopted by the present invention is:
[0058] A drug addict identification device, comprising:
[0059] at least one processor;
[0060] at least one memory for storing at least one program;
[0061] When the at least one program is executed by the at least one processor, the at least one processor implements the above method.
[0062] The third technical solution adopted by the present invention is:
[0063] A computer-readable storage medium stores a program executable by a processor, wherein the program executable by the processor is used to execute the method described above when executed by the processor.
[0064] The beneficial effects of the present invention are: by introducing differential entropy (DE) as the feature representation of EEG signals and combining incremental width learning as a classifier, the present invention can achieve rapid adaptation and efficient training, performs well in processing high-dimensional data and nonlinear classification problems, avoids the high computing resource consumption caused by a large number of parameter training in deep learning, can realize edge computing deployment, and can effectively classify and identify EEG signals. BRIEF DESCRIPTION OF THE DRAWINGS
[0065] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the embodiments of the present invention or the drawings of related technical solutions in the prior art are introduced below. It should be understood that the drawings introduced below are only for the convenience of clearly describing some embodiments of the technical solutions of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.
[0066] Figure 1 is a flowchart of the steps of a method for identifying drug addicts in an embodiment of the present invention;
[0067] Figure 2 It is a schematic diagram of recognition through training detection model in traditional evaluation scheme. DETAILED DESCRIPTION
[0068] The embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and are not to be construed as limitations of the present invention. For the step numbers in the following embodiments, they are only provided for the convenience of explanation, and the order between the steps is not limited in any way, and the execution order of each step in the embodiment can be adaptively adjusted according to the understanding of those skilled in the art.
[0069] In the description of the present invention, it should be understood that descriptions involving orientations, such as up, down, front, back, left, right, etc., and orientations or positional relationships indicated are based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as a limitation on the present invention.
[0070] In the description of the present invention, the meaning of "several" is one or more, the meaning of "more" is two or more, and the meanings of "greater than", "less than", "exceed" and the like are understood as not including the number itself, and the meanings of "above", "below", "within" and the like are understood as including the number itself. If there is a description of the first and the second, it is only used for the purpose of distinguishing the technical features, and cannot be understood as indicating or implying the relative importance or implicitly indicating the number of the indicated technical features or implicitly indicating the order of the indicated technical features. In addition, "and / or" describes the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the related objects before and after are in an "or" relationship.
[0071] In the description of the present invention, unless otherwise clearly defined, terms such as setting, installing, connecting, etc. should be understood in a broad sense, and technicians in the relevant technical field can reasonably determine the specific meanings of the above terms in the present invention based on the specific content of the technical solution.
[0072] See also Figure 2, the subjectivity and inefficiency of traditional evaluation methods still cannot meet actual needs, resulting in inaccurate and inefficient evaluation; and the current EEG-based methods often target individual drug types, lack universality, and require a large number of preprocessing methods to ensure the effectiveness of features. In response to this technical problem, the present invention proposes a method for identifying multiple drug addicts based on EEG, differential entropy and incremental width learning, which covers a variety of drug types from the data aspect. In order to enhance the model's feature extraction of EEG signals and multiple drugs, differential entropy (DE) is introduced as a feature representation of EEG signals. Differential entropy is an effective feature representation method that can capture the dynamic features and complexity in EEG signals and has good distinguishing ability and robustness. Combined with incremental width learning as a classifier, incremental width learning is an efficient machine learning method that learns and updates by gradually adding new data on the basis of the existing model. Unlike traditional deep learning methods, incremental width learning does not require retraining the entire model, but updates parameters incrementally to achieve fast adaptation and efficient training. It performs well in processing high-dimensional data and nonlinear classification problems. In addition, incremental width learning avoids the high computing resource consumption caused by a large number of parameter training in deep learning by adding data and parameters layer by layer. It can realize edge computing deployment and effectively classify and identify EEG signals. Therefore, we try to improve the current drug addict identification method from three aspects: efficiency, accuracy and practicality.
[0073] The present invention can be used in front-line drug enforcement units to test potential drug users, and plays a vital role in drug enforcement and compulsory isolation and detoxification. The addiction assessment technology based on EEG data has penetrated into various related applications due to its objectivity, speed and efficiency, such as quickly identifying the addiction status of drug-related personnel in drug enforcement and improving law enforcement efficiency; analyzing the effects of drugs on the brain in psychological research and providing research data support; and assisting doctors in making accurate diagnoses in medical diagnosis.
[0074] like Figure 1 As shown, this embodiment provides a method for identifying drug addicts, comprising the following steps:
[0075] A method for identifying a drug addict comprises the following steps:
[0076] S1. Obtaining EEG data, which includes EEG data of healthy people and EEG data of drug addicts.
[0077] S2. Preprocess the EEG data to obtain EEG signals.
[0078] EEG data preprocessing is a key step in EEG signal analysis, which aims to extract clean and accurate signals for subsequent analysis. Specifically, the process usually includes steps such as importing raw data, data filtering, data segmentation, baseline correction, and artifact and motion artifact removal.
[0079] In some embodiments, first, the raw EEG data is loaded into the memory by selecting a suitable reading method. Next, the 0.5-45 Hz filtering technique is used to remove noise and unnecessary frequency components in the signal, and then the data is divided into 2 second time periods. Subsequently, a baseline correction is performed to eliminate possible DC offsets, and corresponding algorithms or methods are used to remove artifacts and motion artifacts. Through these steps, high-quality and high-accuracy EEG signals can be obtained, providing a reliable data basis for subsequent signal analysis and feature extraction.
[0080] S3. Use differential entropy to extract dynamic features of the EEG signal to obtain EEG signal features.
[0081] Differential entropy is a statistic used to describe signal complexity and information content, and is often used in EEG signal analysis and feature extraction. Differential entropy reflects the uncertainty of the signal and the degree of change over time, and is an effective EEG signal feature. In some embodiments, the calculation of differential entropy includes the following steps:
[0082] S31, data segmentation: First, the EEG signal is divided into several time periods of equal length to calculate the differential entropy. Since the data has been segmented in step S2, this step can be omitted in step S3.
[0083] S32, signal difference: for each time period of the EEG signal, calculate its first-order difference sequence. As an optional implementation, the difference sequence can be calculated by the difference between the EEG signal values at adjacent time points.
[0084] S33, estimate probability density function: estimate the probability density function according to the differential sequence of the signal. In some optional implementations, the probability density function is usually estimated by using a histogram method or a kernel density estimation method.
[0085] S34. Calculate differential entropy: Use the estimated probability density function and apply the definition formula of differential entropy to calculate. The definition of differential entropy is as follows:
[0086]
[0087] Among them, p(x) represents the probability density function of the signal, and x represents the value of the differential signal. Differential entropy DE represents the amount of information and complexity of the signal differential sequence. The larger the value, the more complex the signal and the richer the information.
[0088] S35, feature extraction: After calculating the differential entropy value of each time period, it is used as one of the features of the EEG signal. As an optional implementation, it is usually combined with other features for analysis and modeling to identify different EEG patterns or determine the state of the EEG signal.
[0089] The application of differential entropy in EEG signal analysis is based on the time domain characteristics and dynamic characteristics of EEG signals. EEG signals are electrophysiological signals of brain neuron activity, which contain rich information, such as the spatiotemporal distribution, frequency characteristics, and time series changes of brain activity. Differential entropy, as a statistic that reflects signal complexity and information volume, can reveal the dynamic change characteristics of EEG signals from a time series perspective, and is therefore closely related to the analysis of EEG signals.
[0090] In EEG signal analysis, differential entropy is often used to study the complexity and uncertainty of EEG signals. EEG signals have different dynamic characteristics in different brain regions and different physiological states. For example, the complexity and amount of information of EEG signals will change in different cognitive tasks, different emotional states or different neurological diseases, which provides important clues for understanding the spatiotemporal dynamics of brain activity.
[0091] In addition, differential entropy can also be used to compare the complexity and information content of different EEG signals. For example, by comparing the differential entropy values of EEG signals between the control group and the experimental group, the differences in brain activity under different cognitive tasks or different neurological disease states can be found. Therefore, differential entropy, as a flexible and effective feature extraction method, has broad application prospects in EEG signal analysis and neuroscience research.
[0092] S4. Input the obtained EEG signal features into incremental width learning for training, and verify the results through ten-fold cross validation.
[0093] The data is input into the incremental width learning for training. Finally, the results are verified through ten-fold cross validation. The verification indicators mainly include accuracy, specificity, sensitivity and ROC / AUC. The incremental width learning system specifically includes the following contents:
[0094] 1) The learning process of width learning:
[0095] Input feature data X, first undergo random mapping to form mapping feature node Z n ≡[Z 1 ,Z 2 ,…,Z n ]; where the i-th group of mapping features Z i The calculation of is shown in formula (1), each set of features contains k mapping features:
[0096]
[0097] The mapping feature node Z obtained by n After the nonlinear mapping function, the enhanced feature node H is formed m ≡[H 1 ,H 2 ,…,H m ]; the jth enhanced feature H j The calculation of is shown in formula (2):
[0098] H j ≡ζ(Z n W hj +β hj ),j=1,2,…,m (2)
[0099] The two features are combined to form a new addiction EEG recognition feature A = [Z|H], and connected to the model output layer Y;
[0100] Since it is known whether each sample has taken drugs, it is only necessary to calculate the connection weight W between the feature layer and the output layer. The calculation formula is as follows:
[0101] W=A -1 Y (3)
[0102] During calculation, ridge regression is used to solve the weight matrix, that is, W is solved by the optimization problem of formula (4):
[0103]
[0104] Among them, σ 1 =σ 2 =v=u=2, the solution is:
[0105] W=(λI+AA T ) -1 A T Y (5)
[0106] The width learning prediction output is:
[0107]
[0108] The confidence level corresponding to the output is obtained by formula (7), which is:
[0109]
[0110] The learning process of data increment width learning:
[0111] For the newly added input data {X a ,Y a}, and calculate the corresponding incremental mapping features and enhanced nodes according to formulas (8) and (9), respectively.
[0112]
[0113]
[0114] Then the new feature layer is:
[0115]
[0116] Output layer weights Updated by the following formula:
[0117]
[0118] in, is the label corresponding to the data Ba;
[0119] The above method is described in detail below in combination with experimental tests.
[0120] In view of the problems of high cost, complex data collection and inability to conduct large-scale screening in the current traditional drug detection methods, and the problems of poor universality and inability to be put into practical application due to the current EEG-based detection methods only targeting single drugs and complex preprocessing processes, as well as the problem that the EEG data sample size used in training is small and cannot cover new drugs and other types of people, this embodiment is based on EEG data combined with differential entropy feature extraction and incremental width learning, which significantly improves the accuracy and real-time performance of detecting multiple drug addicts, and when new drug users or new samples appear, new features can be learned through incremental learning, which is suitable for large-scale screening by law enforcement units. The information of addicts and healthy people in the sample is shown in Tables 1 and 2:
[0121] Table 1
[0122] Number of people Age (years) male female Addicted people 80 37.6(±5.1) 48 32 Healthy controls 70 30.6(±7.3) 36 34
[0123] Table 2
[0124]
[0125] The differential entropy feature can effectively capture the dynamic changes of EEG signals and provide high-resolution time-frequency domain analysis. In actual drug rehabilitation centers, in tests covering multiple types of drugs, the evaluation model has an accuracy rate of 96.429% in identifying addicted individuals, as shown in Table 3. This method solves the problems of complex acquisition and low efficiency in traditional evaluation methods, and also solves the problems of complex preprocessing and poor universality of current EEG-based recognition methods due to single drug targeting.
[0126] Table 3
[0127] Accuracy Incremental Width Learning 96.429%
[0128] This embodiment also provides a drug addict identification device, comprising:
[0129] at least one processor;
[0130] at least one memory for storing at least one program;
[0131] When the at least one program is executed by the at least one processor, the at least one processor implements Figure 1 The method shown.
[0132] A drug addict identification device of this embodiment can execute a drug addict identification method provided by the method embodiment of the present invention, can execute any combination of implementation steps of the method embodiment, and has the corresponding functions and beneficial effects of the method.
[0133] The present application also discloses a computer program product or a computer program, which includes a computer instruction stored in a computer-readable storage medium. A processor of a computer device can read the computer instruction from the computer-readable storage medium, and the processor executes the computer instruction, so that the computer device executes Figure 1 The method shown.
[0134] This embodiment also provides a storage medium storing instructions or programs that can execute a method for identifying drug addicts provided by the method embodiment of the present invention. When the instructions or programs are run, any combination of implementation steps of the method embodiment can be executed, and the corresponding functions and beneficial effects of the method can be obtained.
[0135] In some selectable embodiments, the function / operation mentioned in the block diagram may not occur in the order mentioned in the operation diagram. For example, depending on the function / operation involved, the two boxes shown in succession can actually be executed substantially simultaneously or the boxes can sometimes be executed in reverse order. In addition, the embodiment presented and described in the flow chart of the present invention is provided by way of example, for the purpose of providing a more comprehensive understanding of technology. The disclosed method is not limited to the operation and logic flow presented herein. Selectable embodiments are expected, wherein the order of various operations is changed and the sub-operation of a part for which is described as a larger operation is performed independently.
[0136] In addition, although the present invention is described in the context of functional modules, it should be understood that, unless otherwise specified, one or more of the functions and / or features described may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in separate physical devices or software modules. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the present invention. More specifically, in view of the properties, functions, and internal relationships of the various functional modules in the device disclosed herein, the actual implementation of the module will be understood within the conventional skills of the engineer. Therefore, those skilled in the art can implement the present invention set forth in the claims without excessive experimentation using ordinary techniques. It is also understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the present invention, which is determined by the full scope of the appended claims and their equivalents.
[0137] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.
[0138] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.
[0139] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.
[0140] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.
[0141] In the above description of this specification, the description with reference to the terms "one embodiment / example", "another embodiment / example" or "certain embodiments / examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.
[0142] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.
[0143] The above is a specific description of the preferred implementation of the present invention, but the present invention is not limited to the above embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present invention. These equivalent modifications or substitutions are all included in the scope defined by the claims of this application.
Claims
1. A method for identifying drug addicts, characterized in that: The following steps are involved: Obtain EEG data; Preprocessing the EEG data to obtain EEG signals; Extracting dynamic features of the EEG signal using differential entropy to obtain EEG signal features; The obtained EEG signal features are input into incremental width learning for training, and the results are verified by ten-fold cross validation.
2. A method for identifying drug addicts according to claim 1, characterized in that: The preprocessing includes at least one of data filtering processing, data segmentation processing, baseline correction processing, or artifact and motion artifact removal processing.
3. A method for identifying drug addicts according to claim 1, characterized in that: The preprocessing of the EEG data to obtain an EEG signal includes: The 0.5-45 Hz filtering technology is used to remove noise and unnecessary frequency components in the EEG data, and the filtered EEG data is divided into time periods of preset time lengths.
4. A method for identifying drug addicts according to claim 1, characterized in that: The step of extracting dynamic features of the EEG signal using differential entropy to obtain EEG signal features includes: For each time period of the EEG signal, calculate its first-order difference sequence; Estimate the probability density function of a first-difference sequence; Using the estimated probability density function, calculate the differential entropy; After calculating the differential entropy value of each time period, the differential entropy value is used as the EEG signal feature.
5. A method for identifying drug addicts according to claim 4, characterized in that: The first-order difference sequence is obtained by calculating the difference between the values of the EEG signals at adjacent time points; The probability density function is estimated by using a histogram method or a kernel density estimation method.
6. A method for identifying drug addicts according to claim 4, characterized in that: The calculation formula of the differential entropy is as follows: Where p(x) represents the probability density function of the signal, and x represents the value of the differential signal.
7. A method for identifying drug addicts according to claim 1, characterized in that: The step of inputting the obtained EEG signal features into incremental width learning for training comprises: The learning process of width learning: Input feature data X, first undergo random mapping to form mapping feature node Z n ≡[Z1,Z2,…,Z n ]; where the i-th group of mapping features Z i The calculation of is shown in formula (1): In the formula, is a linear mapping function, W ei is a random weight, β ei is the random bias, n is the number of groups of mapping features; The mapping feature node Z obtained n After the nonlinear mapping function, the enhanced feature node H is formed m ≡[H1,H2,…,H m ]; the jth enhanced feature H j The calculation of is shown in formula (2): H j ≡ζ(Z n W hj +β hj ),j=1,2,…,m (2) Where W hj , β hj is the connection weight, m is the number of enhanced nodes, ζ(Z,W hj ,β hj ) is a nonlinear activation function; the two features are combined to form a new addiction EEG recognition feature A = [Z|H], and connected to the model output layer Y; since it is known whether each sample has taken drugs, it is only necessary to calculate the connection weight W between the feature layer and the output layer, and the calculation formula is as follows: W=A -1 Y (3) During calculation, ridge regression is used to solve the weight matrix, that is, W is solved by the optimization problem of formula (4): Among them, σ1=σ2=v=i=2, and the solution is: W=(λI+AA T ) -1 A T Y (5) In the formula, Y is the true output, I is the unit matrix, λ is the optimization coefficient, and T represents the transpose; The width learning prediction output is: The learning process of data increment width learning: For the newly added input data {X a ,Y a }, and calculate the corresponding incremental mapping features and enhanced nodes according to formulas (7) and (8): Then the new feature layer is: Output layer weights Updated by the following formula: in, Is the label corresponding to the data Ba.
8. A method for identifying drug addicts according to claim 7, characterized in that: The calculation formula for the confidence corresponding to the width learning prediction output Y is as follows: In the formula, represents the output of the sparse width living body recognition network, Represents the output corresponding to the jth category determined by the sparse width living body recognition network.
9. A drug addict identification device, characterized in that: include: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method according to any one of claims 1 to 8.
10. A computer-readable storage medium storing a program executable by a processor, characterized in that: The processor-executable program is used to perform the method according to any one of claims 1 to 8 when executed by the processor.
Citation Information
Patent Citations
Automatic sleep stage division method based on width neural network
CN110859593A
Multi-core width learning-based electroencephalogram emotion classification method and device, medium and equipment
CN113011493A
Methamphetamine addict and normal person classification method based on electroencephalogram signals
CN113040772A
Electroencephalogram fatigue prediction method based on self-weighted increment RVFL network
CN113988123A
Channel screening method, emotion recognition method and system and storage medium
CN114818786A
