TWI810988B - Method of enhancing classification of electroencephalography signals by time-frequency domain channel weighted technique and system thereof - Google Patents

Abstract

A method of enhancing classification of electroencephalography signals by a time-frequency domain channel weighted technique is proposed. The method includes a training step and a testing step. The training step is performed to train a first electroencephalography signal, and the testing step is performed to test and classify a second electroencephalography signal. The first electroencephalography signal and the second electroencephalography signal are corresponding to a training data set and a testing data set, respectively. The training step includes a convolution operation step, a channel weight generating step, a first time-frequency conversion step, a time-frequency weight generating step and a weighted time-frequency generating step. The testing step includes a channel weight calculating step, a second time-frequency conversion step, a time-frequency weight calculating step and a classifying step. Therefore, the method of the present disclosure can extract key features by analyzing the electroencephalography signals of different motor imagery tasks and combine with the time-frequency domain channel weighted technique, so that the types of the motor imagery tasks of the corresponding actions can be effectively distinguished.

Description

Method and system for time-frequency domain channel weighting to strengthen brain wave signal classification

The present invention relates to a method and system for classifying brain wave signals, in particular to a method and system for classifying brain wave signals with time-frequency domain channel weighting enhancement.

Physiological signals will produce different results due to individual differences. In addition, the instability of brain activity and the low signal-to-noise ratio of the signal will limit the performance of motor imagery classification. In the conventional studies on motor imagery, most of the methods did not fully consider the characteristic information of the electroencephalography signal (EEG Signal) in the frequency, time and space domains, and the structure of these models could not effectively extract the Discriminative features, which in turn lead to limited performance in classification. It can be seen that there is currently a lack of a brainwave signal classification method and system in the market that fully considers the time-frequency domain, can effectively extract discriminative features, and improve classification performance. Therefore, relevant companies are looking for solutions.

Therefore, the object of the present invention is to provide a method and system for classifying brainwave signals with time-frequency domain channel weighting enhancement. By analyzing brainwave signals of different motor imagery tasks, key features are extracted from them and combined with time-frequency domain channels. Weighting can effectively distinguish the brain wave form of the corresponding action, and can avoid the problem that the conventional technology cannot effectively extract discriminative features, resulting in limited classification performance.

One embodiment of the method aspect according to the present invention provides a method for time-frequency domain channel weighting enhancement of brain wave signal classification, which includes a training step and a testing step. The training step includes a convolution operation step, a channel weight generation step, a first time-frequency conversion step, a time-frequency weight generation step and a weighted time-frequency generation step. The convolution operation step includes driving the processor to perform convolution operation on the first electroencephalogram signal, and extracting a plurality of first characteristic information of the first electroencephalogram signal. The channel weight generating step includes driving the processor to perform pooling operation on the first feature information, and learning the importance of the first feature information to generate complex channel weight coefficients corresponding to the first feature information. The first time-frequency conversion step includes driving the processor to perform continuous wavelet conversion on the first brain wave signal to generate complex first time-frequency images, these first time-frequency images include the first motion imagery average feature time-frequency image and the second motion Imagine an average feature time-frequency plot. The time-frequency weight generating step includes driving the processor to calculate the difference between the first motor imagery average characteristic time-frequency diagram and the second motor imagery average characteristic time-frequency diagram according to the T test to generate the time-frequency weight. The weighted time-frequency generation step includes driving the processor to multiply the first motor imagery average feature time-frequency graph with the time-frequency weight to generate the first motor imagery weighted feature time-frequency graph, and multiply the second motor imagery average feature time-frequency graph with the time-frequency weight Frequency weights are multiplied to generate a second motor imagery weighted feature time-frequency map. Furthermore, the testing step includes a channel weight calculation step, a second time-frequency conversion step, a time-frequency weight calculation step and a test classification step. The channel weight calculation step includes driving the processor to calculate the second electroencephalogram signal and the channel weight coefficients to generate channel weight signals. The second time-frequency conversion step includes driving the processor to perform continuous wavelet conversion on the channel weight signal to generate a second time-frequency diagram. The time-frequency weight calculation step includes driving the processor to multiply the second time-frequency diagram by the time-frequency weight to generate a weighted test time-frequency diagram. The test classification step includes driving the processor to calculate the weighted test time-frequency map, the first motor imagery weighted feature time-frequency map and the second motor imagery weighted feature time-frequency map to generate a classification result.

In this way, the method of time-frequency domain channel weighting of the present invention to strengthen the classification of brain wave signals can effectively distinguish the corresponding action by analyzing the brain wave signals of different motor imagery tasks, extracting key features, and combining time-frequency domain channel weighting. Brain wave form, and can avoid the problem of limited classification performance caused by the inability of conventional techniques to effectively extract discriminative features.

One embodiment of the structural aspect according to the present invention provides a system for time-frequency domain channel weighting enhanced brain wave signal classification, which includes a storage unit and a processor. The storage unit is used for accessing the first electroencephalogram signal and the second electroencephalogram signal. The processor is connected to the storage unit and receives the first brain wave signal and the second brain wave signal, and the processor is configured to implement the operation including the following steps: a training step and a testing step. The training step includes a convolution operation step, a channel weight generation step, a first time-frequency conversion step, a time-frequency weight generation step and a weighted time-frequency generation step. The convolution operation step includes performing convolution operation on the first electroencephalogram signal, and extracting plural first characteristic information of the first electroencephalogram signal. The channel weight generating step includes performing a pooling operation on the first feature information, and learning the importance of the first feature information to generate complex channel weight coefficients corresponding to the first feature information. The first time-frequency conversion step includes performing continuous wavelet conversion on the first brain wave signal to generate a plurality of first time-frequency images, these first time-frequency images include a first motor imagery average feature time-frequency map and a second motor imagery average feature Time-Frequency Diagram. The time-frequency weight generating step includes calculating the difference between the first motor imagery average characteristic time-frequency diagram and the second motor imagery average characteristic time-frequency diagram according to the T-test to generate the time-frequency weight. The weighted time-frequency generating step comprises multiplying the first motion imagery average feature time-frequency map with the time-frequency weight to generate the first motion imagery weighted feature time-frequency map, and combining the second motion imagery average feature time-frequency map with the time-frequency weight Multiplied to generate a second motor imagery weighted feature time-frequency map. In addition, the testing step includes a channel weight calculation step, a second time-frequency conversion step, a time-frequency weight calculation step and a test classification step. The channel weight calculation step includes calculating the second electroencephalogram signal and the channel weight coefficients to generate channel weight signals. The second time-frequency conversion step includes performing continuous wavelet conversion on the channel weight signal to generate a second time-frequency diagram. The time-frequency weight calculation step includes multiplying the second time-frequency map by the time-frequency weight to generate a weighted test time-frequency map. The test classification step includes calculating the weighted test time-frequency map, the first motor imagery weighted feature time-frequency map and the second motor imagery weighted feature time-frequency map to generate a classification result.

In this way, the time-frequency domain channel weighting system of the present invention enhances brain wave signal classification by analyzing the brain wave signals of different motor imagery tasks, extracting key features from them, and combining time-frequency domain channel weighting, it can effectively distinguish the corresponding action Brain wave form, and can avoid the problem of limited classification performance caused by the inability of conventional techniques to effectively extract discriminative features.

Several embodiments of the present invention will be described below with reference to the drawings. For the sake of clarity, many practical details are included in the following narrative. It should be understood, however, that these practical details should not be used to limit the invention. That is, in some embodiments of the present invention, these practical details are unnecessary. In addition, for the sake of simplifying the drawings, some commonly used structures and elements will be shown in a simple and schematic way in the drawings; and repeated elements may be denoted by the same reference numerals.

In addition, when a certain element (or unit or module, etc.) is "connected" to another element herein, it may mean that the element is directly connected to another element, or it may mean that a certain element is indirectly connected to another element , that is, there are other elements interposed between the element and another element. And when it is stated that an element is "directly connected" to another element, it means that there is no other element interposed between the element and another element. The terms first, second, third, etc. are used to describe different components, and have no limitation on the components themselves. Therefore, the first component can also be called the second component. Moreover, the combination of components/units/circuits in this article is not a combination that is generally known, conventional or conventional in this field. Whether the components/units/circuits themselves are known or not can be used to determine whether the combination relationship is easily recognized by those in the technical field. Usually knowledgeable people do it easily.

Please refer to FIG. 1 . FIG. 1 is a schematic flowchart of a method 100 for enhancing brain wave signal classification with time-frequency domain channel weighting according to a first embodiment of the present invention. The method 100 for enhancing brainwave signal classification with time-frequency domain channel weighting includes a training step S0 and a testing step S2. The training step S0 is used to train the first electroencephalogram signal 110 , and the testing step S2 is used to test and classify the second electroencephalogram signal 120 .

The training step S0 includes a convolution operation step S01, a channel weight generation step S02, a first time-frequency conversion step S03, a time-frequency weight generation step S04, and a weighted time-frequency generation step S05.

The convolution operation step S01 includes driving the processor to perform convolution operation on the first electroencephalogram signal 110 , and extracting a plurality of first characteristic information of the first electroencephalogram signal 110 . Specifically, in the convolution operation step S01 , the processor performs convolution operation on the first electroencephalogram signal 110 by using complex convolution kernels of different sizes to extract the first feature information of different sizes. The first EEG signal 110 belongs to training data. In an embodiment, the sizes of these convolution kernels may be 1×3, 1×5, and 1×11 respectively, but the present invention is not limited thereto.

The channel weight generating step S02 includes driving the processor to perform a pooling operation on the first feature information, and learning the importance of the first feature information to generate the complex channel weight coefficients 112 corresponding to the first feature information. In detail, the channel weight generation step S02 is to correct the first feature information, pool the feature maps on each channel, and then pay more attention to channel features with a large amount of information by learning the importance of different channel features , and suppress the channel features that are not very useful for the current task, and finally obtain the channel weight coefficient 112; in other words, the channel weight coefficient 112 represents the importance of each channel. In one embodiment, the pooling operation may be a global average pooling (Global Average Pooling; GAP) operation, but the invention is not limited thereto.

The first time-frequency conversion step S03 includes driving the processor to perform a continuous wavelet conversion on the first brainwave signal 110 to generate complex first time-frequency images 114, and these first time-frequency images 114 include the average characteristic time-frequency of the first motor imagery Graph 1142 and second motor imagery mean feature time-frequency graph 1144 . In detail, motor imagery refers to imagining and simulating the state of oneself when performing sports in the brain. The first motor imagery average feature time-frequency graph 1142 represents the average of a plurality of first motor imagery feature time-frequency graphs, which may correspond to a left-hand motor imagery task; the second motor imagery average feature time-frequency graph 1144 represents a plurality of second motor imagery The average of characteristic time-frequency maps, which can correspond to a right-hand motor imagery task. The continuous wavelet transform may be a Morlet wavelet transform, which can effectively extract feature information from the transient first EEG signal 110 through local transformation of time and frequency. After the first brainwave signal 110 is converted to the time-frequency domain, it can be drawn into a two-dimensional time-frequency diagram, so as to observe the energy level changes at a certain frequency and within a certain period of time in this experiment, and also through observing multiple experiments. Estimate the event-related synchronous brain wave rhythm (Event-Related Synchronization; ERS) and event-related asynchronous brain wave rhythm (Event-Related Desynchronization; ERD), which is beneficial to the analysis of motor imagery, as well as the changes of mu wave and beta wave. The continuous wavelet transform conforms to the following formula (1): (1). in, Represents the wavelet basis function generated by the mother wavelet by stretching a and translating b coefficients . representing the input first brain wave signal 110, represents the result after continuous wavelet transformation, s represents the motor imagery state of different tasks, and C represents the channel.

The time-frequency weight generating step S04 includes driving the processor to calculate a difference between the first motor imagery average characteristic time-frequency diagram 1142 and the second motor imagery average characteristic time-frequency diagram 1144 according to a T test to generate a time-frequency weight 116 . In detail, the T-test can be an independent samples T-test (Independent Samples T-Test). The time-frequency weight generation step S04 includes driving the processor to calculate an average value and a standard deviation of any one of the first motor imagery average feature time-frequency graph 1142 and the second motor imagery average feature time-frequency graph 1144; and driving the processor according to The T-test calculates the mean value and standard deviation to obtain the time-frequency weight 116 .

The average value was calculated based on the complex motor imagery task status, the complex total number of trials, the complex channels and the complex time points. The average value is expressed as , the state of these motor imagery tasks is denoted as s , and the total number of trials is denoted as , these channels are denoted as C , and these time points are denoted as b . The average value conforms to the following formula (2): (2).

The standard deviation is calculated based on the mean value, the motor imagery task status, the total number of trials, the channels and the time points. The standard deviation is expressed as , and conform to the following formula (3): (3).

The time-frequency weight 116 is calculated and obtained according to the first motor imagery task state, the second motor imagery task state, the total number of first motor imagery tests, the total number of second motor imagery tests, these channels and these time points. The time-frequency weight 116 is expressed as , the first motor imagery task state is denoted as s ₁ , the second motor imagery task state is denoted as s ₂ , and the total number of first motor imagery trials is denoted as , the total number of second motor imagery trials is expressed as , these channels are denoted as C , these time points are denoted as b , and the time-frequency weight 116 conforms to the following formula (4): (4). obtained above Represents the degree of difference between the two groups of motor imagery tasks. The present invention can observe the differences of different time points b and different frequencies between two groups of data (for example: The point of the local maximum represents the maximum difference between the motor imagery state of the left hand and the right hand at the time point b ), and this value is used as the time-frequency weight.

The weighted time-frequency generation step S05 includes driving the processor to multiply the first motion imagery average feature time-frequency graph 1142 with the time-frequency weight 116 to generate the first motion imagery weighted feature time-frequency graph 1182, and multiply the second motion imagery average feature time-frequency graph 1142 The frequency map 1144 is multiplied by the time-frequency weights 116 to generate a second motor imagery weighted feature time-frequency map 1184 . The first motor imagery weighted feature time-frequency map 1182 corresponds to the left-hand motor imagery task, and the second motor imagery weighted feature time-frequency map 1184 corresponds to the right-hand motor imagery task.

The test step S2 includes a channel weight calculation step S21, a second time-frequency conversion step S22, a time-frequency weight calculation step S23, and a test classification step S24.

The channel weight calculation step S21 includes driving the processor to calculate the second EEG signal 120 and the channel weight coefficients 112 to generate the channel weight signal 122 . The second EEG signal 120 belongs to test data, and the channel weight coefficient 112 comes from the channel weight generation step S02 of the training step S0.

The second time-frequency conversion step S22 includes driving the processor to perform continuous wavelet conversion on the channel weight signal 122 to generate a second time-frequency diagram 124 . The continuous wavelet transform can be the same as the formula (1) in the first time-frequency transform step S03, and the details will not be repeated here.

The time-frequency weight calculation step S23 includes driving the processor to multiply the second time-frequency map 124 by the time-frequency weight 116 to generate a weighted test time-frequency map 126 . The time-frequency weight 116 comes from the time-frequency weight generating step S04 of the training step S0; in other words, the weighted test time-frequency diagram 126 is a T-statistically weighted time-frequency diagram.

The test classification step S24 includes driving the processor to calculate the weighted test time-frequency map 126 , the first motor imagery weighted feature time-frequency map 1182 and the second motor imagery weighted feature time-frequency map 1184 to generate the classification result 128 . Specifically, in the test classification step S24, the processor calculates the weighted test time-frequency diagram 126, the first motor imagery weighted feature time-frequency map 1182 and the second motor imagery weighted feature time-frequency map 1184 to obtain the two correlation coefficients , , and compare the two correlation coefficients , And produce classification results 128, the two correlation coefficients , Either of them conforms to the following formula (5): (5). in, Represents the two correlation coefficients , , Represents a total number of trials, s represents the complex motor imagery task state, x represents one of the first motor imagery weighted characteristic time-frequency diagram 1182 and the second motor imagery weighted characteristic time-frequency diagram 1184, y represents the weighted test time-frequency diagram 126, i stands for 1~ One of the positive integers. For example, if x represents the first motor imagery weighted feature time-frequency map 1182 (corresponding to the left hand motor imagery), then Represents the correlation coefficient ; If x represents the second motor imagery weighted feature time-frequency map 1184 (corresponding to right-hand motor imagery), then Represents the correlation coefficient . Bi-correlation coefficient , Between the weighted test time-frequency diagram 126 and the average time-frequency diagram of left and right hand motor imagery (i.e. the first motor imagery average characteristic time-frequency diagram 1142, the second motor imagery average characteristic time-frequency diagram 1144) used to evaluate each test respectively degree of relevance. Calculate the weighted test time-frequency diagram 126 of a test and the average time-frequency diagram of left-hand motor imagery (i.e. the first motor imagery average characteristic time-frequency diagram 1142), and the correlation coefficient can be obtained ; The weighted test time-frequency diagram 126 of a test and the average time-frequency diagram of right-hand motor imagery (i.e. the second motor imagery average characteristic time-frequency diagram 1144) are calculated to obtain the correlation coefficient . If the correlation coefficient greater than or equal to the correlation coefficient , then the classification result 128 represents that the test is predicted to be left-handed motor imagery; on the contrary, if the correlation coefficient less than the correlation coefficient , then the classification result 128 represents that the test is predicted to be right-hand motor imagery.

Correlation coefficient of formula (5) Represent the time-frequency diagram weighted by T statistics (i.e. weighted test time-frequency diagram 126) and the average time-frequency diagram (i.e. the first weighted feature time-frequency diagram 1182, the first weighted characteristic time-frequency diagram 1182) of left and right hand motion imagery weighted by T statistics in a test The degree of correlation between the two motor imagery weighted feature time-frequency diagrams 1184). In this way, the present invention can use the correlation coefficient To observe which type of motor imagery task is closely related to this experiment, and speculate on its classification. In addition, by analyzing the brain wave signals of different motor imagery tasks and extracting key features from them, the brain wave forms of corresponding actions can be effectively distinguished.

Please refer to FIG. 1 and FIG. 2 together. FIG. 2 is a schematic diagram of a system 200 for enhancing brain wave signal classification with time-frequency domain channel weighting according to a second embodiment of the present invention. The system 200 for enhancing brain wave signal classification with time-frequency domain channel weighting includes a storage unit 210 and a processor 220 . The storage unit 210 is used for accessing the first electroencephalogram signal 110 and the second electroencephalogram signal 120 . The processor 220 is connected to the storage unit 210 and receives the first electroencephalogram signal 110 and the second electroencephalogram signal 120. The processor 220 is configured to implement the method 100 for time-frequency domain channel weighting and strengthening electroencephalogram signal classification in FIG. 1, that is Processor 220 is configured to perform convolution operation 222 , pooling operation 224 , continuous wavelet transform 226 and T-test 228 . The storage unit 210 may include a random access memory (Random Access Memory; RAM) or other types of dynamic storage devices capable of storing information and instructions for execution by the processor 220 . The processor 220 may include any type of processor, microprocessor, or Field Programmable Gate Array (FPGA) capable of compiling and executing instructions. Processor 220 may comprise a single device (eg, a single core) or a group of devices (eg, multiple cores). In this way, the system 200 of the present invention for time-frequency domain channel weighting to strengthen the classification of brain wave signals can extract key features from the brain wave signals of different motor imagery tasks and combine them with time-frequency domain channel weighting to effectively distinguish corresponding actions. It can avoid the problem that the conventional technology cannot effectively extract the features with discriminative power, which leads to the limitation of classification performance.

It can be seen from the above embodiments that the present invention has the following advantages: First, the feature information can be corrected, and the feature maps on each channel can be pooled, and then by learning the importance of different channel features, further attention can be paid to information-heavy Channel features, and suppress channel features that are not very useful for the current task, and then effectively extract discriminative features. Second, the correlation coefficient can be used to observe which types of motor imagery tasks are closely related to each test, and to speculate on its classification. Third, by analyzing the brain wave signals of different motor imagery tasks, key features can be extracted from them, combined with time-frequency domain channel weighting, the brain wave form of the corresponding action can be effectively distinguished, and the conventional technology can not effectively extract the discriminative features. The characteristics of power lead to the problem of limited classification efficiency.

Although the present invention has been disclosed above in terms of implementation, it is not intended to limit the present invention. Anyone skilled in this art can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the protection of the present invention The scope shall be defined by the appended patent application scope.

100: Time-frequency domain channel weighting method to strengthen the classification of brain wave signals 110: The first brain wave signal 112: Channel weight coefficient 114: The first time-frequency graph 1142: The first motor imagery average feature time-frequency graph 1144: The second motor imagery Average feature time-frequency diagram 116: time-frequency weight 1182: first motor imagery weighted feature time-frequency graph 1184: second motor imagery weighted feature time-frequency graph 120: second brain wave signal 122: channel weight signal 124: second time-frequency Figure 126: Weighted Test Time-Frequency Figure 128: Classification Results : correlation coefficient S0: training step S01: convolution operation step S02: channel weight generation step S03: first time-frequency conversion step S04: time-frequency weight generation step S05: weighted time-frequency generation step S2: testing step S21: channel weight calculation Step S22: Second Time-Frequency Conversion Step S23: Time-Frequency Weight Calculation Step S24: Test Classification Step 200: Time-Frequency Domain Channel Weighting Enhanced EEG Signal Classification System 210: Storage Unit 220: Processor 222: Convolution Operation 224: Pooling Operation 226: Continuous Wavelet Transformation 228: T Test

FIG. 1 is a schematic flow chart showing the method for time-frequency domain channel weighting and strengthening of brain wave signal classification according to the first embodiment of the present invention; and FIG. 2 is a schematic diagram of a system for time-frequency domain channel weighting and enhanced brain wave signal classification according to the second embodiment of the present invention.

100: Time-Frequency Domain Channel Weighting Enhancement Method for EEG Signal Classification

110: The first brain wave signal

112: channel weight coefficient

114: The first time-frequency diagram

1142: Time-Frequency Diagram of Average Feature of First Motor Imagery

1144: Time-Frequency Diagram of Average Feature of Second Motor Imagery

116: Time-frequency weight

1182: Time-Frequency Map of the First Motion Imagery Weighted Feature

1184: Time-Frequency Diagram of Second Motor Imagery Weighted Feature

120: The second brain wave signal

122: Channel weight signal

124: The second time-frequency diagram

126: Weighted test time-frequency diagram

128: Classification result

r _s : correlation coefficient

S0: training step

S01: Convolution operation steps

S02: Channel weight generation steps

S03: the first time-frequency conversion step

S04: Time-frequency weight generation steps

S05: weighted time-frequency generation steps

S2: Test steps

S21: Channel weight calculation steps

S22: the second time-frequency conversion step

S23: Time-frequency weight calculation steps

S24: Test classification step

Claims (10)

A method for weighting and enhancing the classification of brainwave signals in the time-frequency domain, comprising the following steps: A training step, including: A convolution operation step, including driving a processor to perform a convolution operation on a first electroencephalogram signal, and extracting a plurality of first characteristic information of the first electroencephalogram signal; A channel weight generation step, including driving the processor to perform a pooling operation on the first feature information, and learning the importance of the first feature information to generate complex channel weight coefficients corresponding to the first feature information; A first time-frequency conversion step, including driving the processor to perform a continuous wavelet conversion on the first brain wave signal to generate a plurality of first time-frequency images, the first time-frequency images include a first motor imagery average feature time A frequency map and a second motor imagery average characteristic time-frequency map; A time-frequency weight generation step includes driving the processor to calculate a difference between the first motor imagery average feature time-frequency map and the second motor imagery average feature time-frequency map according to a T test to generate a time-frequency weight; and A weighted time-frequency generation step, including driving the processor to multiply the first motion imagery average feature time-frequency graph with the time-frequency weight to generate a first motion imagery weighted feature time-frequency graph, and the second motion imagery multiplying the average feature time-frequency map by the time-frequency weight to generate a second motor imagery weighted feature time-frequency map; and A test procedure, including: A channel weight calculation step, including driving the processor to calculate a second electroencephalogram signal and the channel weight coefficients to generate a channel weight signal; a second time-frequency conversion step, including driving the processor to perform the continuous wavelet conversion on the channel weight signal to generate a second time-frequency diagram; A time-frequency weight calculation step, including driving the processor to multiply the second time-frequency diagram by the time-frequency weight to generate a weighted test time-frequency diagram; and A test classification step includes driving the processor to calculate the weighted test time-frequency map, the first motion imagery weighted feature time-frequency map and the second motion imagery weighted feature time-frequency map to generate a classification result. The method of time-frequency domain channel weighting and strengthening of brain wave signal classification as described in claim 1, wherein, In the convolution operation step, the processor uses complex convolution kernels of different sizes to perform the convolution operation on the first electroencephalogram signal, so as to extract the first feature information of different sizes; and The continuous wavelet is transformed into a Morlet (Morlet) wavelet transform, the first motion imagery average feature time-frequency map and the first motion imagery weighted feature time-frequency map correspond to a left hand motion imagery task, the second motion imagery average feature time-frequency map The frequency map and the second motor imagery weighted feature time-frequency map correspond to a right-hand motor imagery task, the pooling operation is a global average pooling (Global Average Pooling; GAP) operation, and the T-test is an independent sample T-test. The method for time-frequency domain channel weighting enhancement of brain wave signal classification as described in Claim 1, wherein the time-frequency weight generating step further includes: driving the processor to calculate an average value and a standard deviation of any one of the first motor imagery mean feature time-frequency map and the second motor imagery mean feature time-frequency map; and The processor is driven to calculate the average value and the standard deviation according to the T test to obtain the time-frequency weight. The method for time-frequency domain channel weighting enhancement of brainwave signal classification as described in claim 3, wherein in the time-frequency weight generation step, the average value is based on complex motor imagery task status, complex total number of trials, complex channels and complex time The point calculation is obtained, and the average value is expressed as , the state of these motor imagery tasks is denoted as s , and the total number of trials is denoted as , the channels are denoted as C , and the time points are denoted as b , the average value conforms to the following formula: ; The standard deviation is calculated based on the average value, the motor imagery task states, the total number of trials, the channels and the time points, and the standard deviation is expressed as , and conform to the following formula: ; and the time-frequency weight is calculated based on a first motor imagery task state, a second motor imagery task state, a first motor imagery test total number, a second motor imagery test total number, these channels and these time points , the time-frequency weight is expressed as , the first motor imagery task state is denoted as s ₁ , the second motor imagery task state is denoted as s ₂ , and the total number of first motor imagery trials is denoted as , the total number of second motor imagery trials is expressed as , these channels are denoted as C , and these time points are denoted as b , the time-frequency weight conforms to the following formula: . The method for time-frequency domain channel weighting enhancement of brain wave signal classification as described in Claim 1, wherein in the test classification step, the processor calculates the weighted test time-frequency map, the first weighted feature time-frequency map of motor imagery and The second motor imagery weighted characteristic time-frequency diagram obtains two correlation coefficients, and compares the two correlation coefficients to produce the classification result, and each of the two correlation coefficients meets the following formula: ; in, Represents each of the two correlation coefficients, Represents a total number of trials, s represents the complex motor imagery task state, x represents one of the first motor imagery weighted feature time-frequency graph and the second motor imagery weighted feature time-frequency graph, y represents the weighted test time-frequency graph, i stands for 1~ One of the positive integers. A time-frequency domain channel weighted enhanced brain wave signal classification system, including: a storage unit for accessing a first electroencephalogram signal and a second electroencephalogram signal; and A processor, connected to the storage unit and receiving the first electroencephalogram signal and the second electroencephalogram signal, the processor is configured to perform operations comprising the following steps: A training step, including: A convolution operation step, including performing a convolution operation on the first electroencephalogram signal, and extracting a plurality of first characteristic information of the first electroencephalogram signal; A channel weight generation step includes performing a pooling operation on the first feature information, and learning the importance of the first feature information to generate complex channel weight coefficients corresponding to the first feature information; A first time-frequency conversion step includes performing a continuous wavelet conversion on the first electroencephalogram signal to generate complex first time-frequency images, these first time-frequency images include a first motion imagery average characteristic time-frequency image and a The average characteristic time-frequency diagram of the second motor imagery; A time-frequency weight generating step, comprising calculating a difference between the first motor imagery average feature time-frequency map and the second motor imagery average feature time-frequency map according to a T test to generate a time-frequency weight; and A weighted time-frequency generating step, comprising multiplying the first motion imagery average feature time-frequency graph by the time-frequency weight to generate a first motion imagery weighted feature time-frequency graph, and multiplying the second motion imagery average feature time-frequency graph multiplying the map with the time-frequency weight to generate a second motor imagery weighted feature time-frequency map; and A test procedure, including: A channel weight calculation step, including calculating the second electroencephalogram signal and the channel weight coefficients to generate a channel weight signal; a second time-frequency conversion step comprising performing the continuous wavelet conversion on the channel weight signal to generate a second time-frequency diagram; a time-frequency weight calculation step comprising multiplying the second time-frequency diagram by the time-frequency weight to generate a weighted test time-frequency diagram; and A test classification step includes calculating the weighted test time-frequency map, the first motion imagery weighted feature time-frequency map and the second motion imagery weighted feature time-frequency map to generate a classification result. The time-frequency domain channel weighting system for enhancing brain wave signal classification as described in Claim 6, wherein, In the convolution operation step, the processor uses complex convolution kernels of different sizes to perform the convolution operation on the first electroencephalogram signal, so as to extract the feature information of different sizes; and The continuous wavelet is transformed into a Morlet (Morlet) wavelet transform, the first motion imagery average feature time-frequency map and the first motion imagery weighted feature time-frequency map correspond to a left hand motion imagery task, the second motion imagery average feature time-frequency map The frequency map and the second motor imagery weighted feature time-frequency map correspond to a right-hand motor imagery task, the pooling operation is a global average pooling (Global Average Pooling; GAP) operation, and the T-test is an independent sample T-test. The time-frequency domain channel weighting system for enhancing brainwave signal classification as described in Claim 6, wherein the time-frequency weight generating step further includes: driving the processor to calculate an average value and a standard deviation of any one of the first motor imagery mean feature time-frequency map and the second motor imagery mean feature time-frequency map; and The processor is driven to calculate the average value and the standard deviation according to the T test to obtain the time-frequency weight. The time-frequency domain channel weighting system for strengthening brain wave signal classification as described in Claim 8, wherein in the time-frequency weight generation step, the average value is based on complex motor imagery task status, complex total number of trials, complex channels and complex time The point calculation is obtained, and the average value is expressed as , the state of these motor imagery tasks is denoted as s , and the total number of trials is denoted as , the channels are denoted as C , and the time points are denoted as b , the average value conforms to the following formula: ; The standard deviation is calculated based on the average value, the motor imagery task states, the total number of trials, the channels and the time points, and the standard deviation is expressed as , and conform to the following formula: ; and the time-frequency weight is calculated based on a first motor imagery task state, a second motor imagery task state, a first motor imagery test total number, a second motor imagery test total number, these channels and these time points , the time-frequency weight is expressed as , the first motor imagery task state is denoted as s ₁ , the second motor imagery task state is denoted as s ₂ , and the total number of first motor imagery trials is denoted as , the total number of second motor imagery trials is expressed as , these channels are denoted as C , and these time points are denoted as b , the time-frequency weight conforms to the following formula: . The system for time-frequency domain channel weighting and enhanced brain wave signal classification as described in Claim 6, wherein in the test classification step, the processor calculates the weighted test time-frequency map, the first weighted feature time-frequency map of motor imagery and The second motor imagery weighted characteristic time-frequency diagram obtains two correlation coefficients, and compares the two correlation coefficients to produce the classification result, and each of the two correlation coefficients meets the following formula: ; in, Represents each of the two correlation coefficients, Represents a total number of trials, s represents the complex motor imagery task state, x represents one of the first motor imagery weighted feature time-frequency graph and the second motor imagery weighted feature time-frequency graph, y represents the weighted test time-frequency graph, i stands for 1~ One of the positive integers.

Images