Brain-computer interface cap electroencephalogram signal classification method and system for rehabilitation training

Abstract

The invention relates to the field of computers, in particular to a method and a system for classifying electroencephalogram signals in a rehabilitation training process, aiming at the characteristic of low resolution of the electroencephalogram signals, the electroencephalogram signals are firstly preprocessed, and the preprocessing specifically comprises filtering processing, reference correction processing, filling processing and further filtering processing; interference signals in the electroencephalogram signals are removed as much as possible; then the electroencephalogram signals are converted into the time-frequency two-dimensional images, the time-frequency two-dimensional images are classified through the support vector machine classifier model, and the accuracy of electroencephalogram signal classification can be improved.

Description

Technical Field

[0001] The present invention relates to the field of computers, and in particular to a method and system for classifying electroencephalogram (EEG) signals during rehabilitation training. Background Art

[0002] Nowadays, brain-computer interface technology has become an emerging technology that can convert human intentions into control signals, allowing disabled people to interact with the external environment without any real movement. The brain-computer interface system extracts the corresponding electrical signals generated by the brain through signal acquisition equipment, and connects them to external devices after processing, so that it can interact with the outside world. Modern brain neuroscience has proven that people's different psychological states, emotions, and thoughts will affect the changes in EEG signals. Therefore, it is feasible to control equipment through EEG signals to conduct rehabilitation training for patients. At present, brain-computer interface technology has been widely studied and applied in rehabilitation training.

[0003] For example, a Chinese invention patent (CN104951797A) discloses an ELM motor imagery EEG classification method based on AR coefficient space. First, the motor imagery EEG signal of a single channel is taken as a random signal and fitted using a p-order AR model; then the Burg algorithm is used to solve the unknown coefficients a1, a2, …, ap of the AR model; then the p-order AR coefficients of the m-channel motor imagery EEG signals are combined into an AR coefficient vector, which is embedded in the original optimization problem of ELM network parameter training to solve the optimal external weight β and construct an ELM classification algorithm based on AR coefficient space.

[0004] However, when classifying EEG signals, the above method cannot achieve satisfactory EEG signal classification results because the EEG signals are weak signals and there are many interference signals. Therefore, the prior art urgently needs a technical solution for classifying EEG signals of a brain-computer interface cap with high accuracy for rehabilitation training. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a method and system for classifying EEG signals during rehabilitation training in order to improve the accuracy of EEG signal classification in view of the deficiencies of the above technical solutions.

[0006] The technical solution adopted by the present invention is:

[0007] A method for classifying electroencephalogram signals of a brain-computer interface cap for rehabilitation training comprises the following steps:

[0008] Step 1: Collect EEG signals through the brain-computer interface cap;

[0009] Step 2: Preprocess the EEG signal collected in step 1;

[0010] Step 3: Convert the preprocessed EEG signal into a time-frequency two-dimensional image;

[0011] Step 4: Use a support vector machine classifier model to classify the time-frequency two-dimensional image.

[0012] Specifically, in order to make the electrodes of the brain-computer interface cap in close contact with human skin, an automatic tightener is arranged next to the electrodes of the brain-computer interface cap, and the automatic tightener is used to make the electrodes of the brain-computer interface cap in close contact with human skin.

[0013] Furthermore, the automatic tightener includes a pressure sensor, a motor, and a microcontroller; the motor is arranged on the rear edge of the brain-computer interface cap and is connected to the microcontroller, an automatic tightening line is wound on the motor output shaft, a free end of the automatic tightening line is connected to the wedge block through the automatic tightener line, the wedge block is arranged in a transverse channel of the mounting block, the mounting block is arranged on the inner wall of the brain-computer interface cap, and a vertical channel is also arranged on the mounting block, the vertical channel is connected to the transverse channel, and an electrode push rod is arranged in the vertical channel, the top of the electrode push rod is arranged as an inclined surface and contacts the inclined surface of the wedge block, and an electrode is arranged at the bottom thereof, and the pressure sensor is arranged on the motor push rod;

[0014] The pressure sensor is used to detect the contact pressure between the electrode and the human body, and the microcontroller is used to determine whether the contact pressure value is less than the rated pressure value, and when it is less than the rated pressure value, control the motor to enter the working state so that the motor drives the electrode to be in close contact with the human body. Figures 2A-2D shown.

[0015] Furthermore, the wedge block is an I-shaped structure, the vertical plates at both ends of which are located outside the mounting block, and a compression spring is sleeved on the wedge block, which is located outside the mounting block and connected to the side wall of the mounting block.

[0016] The motor winds up the automatic tightener wire, so that the wedge-shaped block pushes the electrode push rod downward and presses on the scalp. The pressure sensor is used to collect pressure signals.

[0017] In the above technical solution, the preprocessing of the EEG signal collected in step 1 specifically includes:

[0018] Step 2.1: Filter the EEG signal;

[0019] A high-pass filter is used to filter out high-frequency interference signals in the EEG signal, and then a band-pass filter is used to perform band-pass filtering on the EEG signal to obtain a required frequency band range;

[0020] Step 2.2: performing baseline correction on the EEG signal processed by S2.1;

[0021] Firstly, the reference values ​​of the EEG signals of multiple channels are determined, and then the multiple reference values ​​are averaged, and then the average value is subtracted from the effective data of the EEG signals of each channel, and the reference correction is realized by using the difference average method;

[0022] Step 2.3: Filling the data of the damaged channel in the EEG signal processed by S2.2;

[0023] The EEG data of the damaged track is filled with the EEG signal data of the adjacent channels;

[0024] Step 2.4: Use principal component analysis to further filter the EEG signal processed in step 2.3 to filter out eye movement interference signals and heart movement interference signals.

[0025] In the above technical solution, the support vector machine classifier model is used to classify the time-frequency two-dimensional image, specifically including:

[0026] Step 4.1: Build a support vector machine classifier;

[0027] Specifically, the support vector machine classifier is:

[0028]

[0029] Wherein, w is the characteristic matrix of the input signal, x is the input signal, and x i is the i-th eigenvector, y i For category, a i is the intermediate quantity, b is the coefficient, K(x i , x) is the kernel function of the classifier, and n represents the number of input signals;

[0030] Step 4.2: Design the structure of the classifier model;

[0031] In the classifier structure design of the present invention, a plurality of classifiers are used to form a classifier group, each classifier is run separately to perform classification prediction, and then the final result is determined by voting based on the classification results of all classifiers;

[0032] Furthermore, the present invention uses 6 classifiers to form the classifier model;

[0033] Step 4.3: Find the optimal parameters through grid search method;

[0034] The present invention determines the optimal parameters by performing grid search on two parameters, the penalty factor coefficient and the kernel function parameter coefficient;

[0035] Step 4.4: inputting the time-frequency two-dimensional image into the support vector machine classifier model to obtain a classification result;

[0036] Specifically, the classification results include: left hand movement, right hand movement.

[0037] According to another aspect of the present invention, the present invention also includes a system for classifying EEG signals during rehabilitation training, the system adopts the above-mentioned method for classifying EEG signals of a brain-computer interface cap for rehabilitation training, and includes:

[0038] An EEG signal acquisition module, used to collect EEG signals through a brain-computer interface cap;

[0039] A preprocessing module, used for preprocessing the EEG signals collected by the EEG signal collection module;

[0040] A signal conversion module converts the preprocessed EEG signal into a time-frequency two-dimensional image;

[0041] The classification module uses a support vector machine classifier model to classify the time-frequency two-dimensional image.

[0042] According to another aspect of the present invention, a computer-readable storage medium is further included, wherein a program is stored on the computer-readable storage medium, and the program is executed by a processor to implement the above-mentioned method for classifying EEG signals during rehabilitation training.

[0043] Beneficial effects of the present invention:

[0044] In view of the characteristics of difficult volume and low resolution of EEG signals, the present invention converts EEG signals into time-frequency two-dimensional images, and classifies the time-frequency two-dimensional images through a support vector machine classifier model. The above classification method can effectively improve the accuracy of EEG signal classification. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0046] Figure 1 A flowchart of a method for classifying EEG signals during rehabilitation training provided in an embodiment of the present application;

[0047] Figure 2A This is a schematic diagram of the effect of wearing a brain-computer interface cap according to an embodiment of the present application;

[0048] Figure 2B This is a schematic diagram of the side view effect of wearing a brain-computer interface cap according to an embodiment of the present application;

[0049] Figure 2CThis is a partial cross-sectional view of a person wearing a brain-computer interface cap according to an embodiment of the present application;

[0050] Figure 2D This is a partial cross-sectional enlarged view of a brain-computer interface cap worn in an embodiment of the present application;

[0051] Figure 3 A flowchart of preprocessing collected EEG signals provided in an embodiment of the present application;

[0052] Figure 4 A flow chart of classifying a time-frequency two-dimensional image using a support vector machine classifier model provided in an embodiment of the present application;

[0053] Figure 5 A schematic diagram of an EEG signal classification system for a rehabilitation training process provided in an embodiment of the present application.

[0054] The labels are:

[0055] 1-Brain-computer interface cap

[0056] 2- Microcontroller

[0057] 3-Motor

[0058] 4-Pressure sensor

[0059] 5-Electrode

[0060] 6-Automatic tensioner line

[0061] 7-Wedge Block

[0062] 8-Electrode push rod DETAILED DESCRIPTION

[0063] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0064] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] Example 1

[0066] In order to achieve the above purpose, Figure 1As shown, the EEG signal classification method during rehabilitation training includes the following steps:

[0067] Step 1: Collect EEG signals through the brain-computer interface cap;

[0068] Specifically, in order to make the electrodes of the brain-computer interface cap in close contact with human skin, an automatic tightener is arranged next to the electrodes of the brain-computer interface cap, and the automatic tightener is used to make the electrodes of the brain-computer interface cap in close contact with human skin.

[0069] Furthermore, if Figures 2A-2D As shown, the automatic tightener includes a pressure sensor, a motor, and a microcontroller; the motor is arranged on the rear edge of the brain-computer interface cap and is connected to the microcontroller, an automatic tightening wire is wound on the motor output shaft, the free end of the automatic tightening wire is connected to the wedge block through the automatic tightener wire, the wedge block is arranged in the transverse channel of the mounting block, the mounting block is arranged on the inner wall of the brain-computer interface cap, and the mounting block is also provided with a vertical channel, the vertical channel is connected to the transverse channel, and an electrode push rod is arranged in the vertical channel, the top of the electrode push rod is arranged as an inclined surface and contacts with the inclined surface of the wedge block, and an electrode is arranged at the bottom thereof, and the pressure sensor is arranged on the motor push rod.

[0070] The wedge block is an I-shaped structure, the vertical plates at both ends of which are located outside the mounting block, and a compression spring is sleeved on the wedge block, which is located outside the mounting block and connected to the side wall of the mounting block.

[0071] The pressure sensor is used to detect the contact pressure between the electrode and the human body, and the microcontroller is used to determine whether the contact pressure value is less than the rated pressure value, and when it is less than the rated pressure value, the motor is controlled to enter the working state, so that the motor drives the electrode to be in close contact with the human body. The motor winds the automatic tightener line, so that the wedge-shaped block pushes the electrode push rod down and presses on the scalp. The pressure sensor is used to collect pressure signals.

[0072] Step 2: preprocessing the EEG signal collected in step 1;

[0073] The EEG signal is a very weak bioelectric signal. Due to the limitation of the signal acquisition equipment, the EEG signals collected are usually around 40 to 110 microvolts, resulting in a low signal-to-noise ratio of the EEG signal. In addition, the signal is very susceptible to the influence of the subject's state and the surrounding environment during acquisition. For example, the noise generated by eye movement and muscles, as well as the background noise of the sensor and the surrounding environment, can interfere with the EEG signal. Therefore, this embodiment sets a preprocessing process to filter out interference signals in the EEG signal as much as possible.

[0074] Specifically, Figure 3 As shown, preprocessing the EEG signal collected in step 1 specifically includes:

[0075] Step 2.1: filtering the EEG signal;

[0076] Specifically, a high-pass filter is used to filter out high-frequency interference signals in the EEG signal, and then a band-pass filter is used to perform band-pass filtering on the EEG signal to obtain a required frequency band range;

[0077] Step 2.2: performing baseline correction on the EEG signal;

[0078] During the EEG signal acquisition process, the collected data sometimes shows an upward deviation trend. When processing the signal, due to the influence of the upward deviation of the data, the amplitude of the EEG signals collected by different electrodes may be different. Therefore, it is necessary to perform a baseline correction on the EEG signal so that the EEG signals collected by different electrodes have the same baseline.

[0079] Specifically, the reference value correction of the EEG signal specifically includes: first determining the reference values ​​of the EEG signals of multiple channels, then performing an average operation on the multiple reference values, then subtracting the average value from the effective data of the EEG signal of each channel, and implementing the reference correction by using the difference average method;

[0080] Step 2.3: Fill the EEG signal data of the damaged channel;

[0081] Since the spatial resolution of EEG signals is low, the EEG signals of each channel will be similar to the EEG signal waveforms of adjacent channels. Therefore, when the collected data is damaged and the data does not meet the requirements, the EEG signal data of the adjacent channels are used to fill the EEG data of the damaged track.

[0082] Step 2.4: further filtering the EEG signal using principal component analysis;

[0083] Eye movement signals and heart movement signals will interfere with EEG signals. Conventional filtering methods are not very effective in filtering out eye movement signals and heart movement signals. Therefore, this embodiment uses principal component analysis to filter the EEG signals to filter out eye movement interference signals and heart movement interference signals.

[0084] Step 3: Convert the preprocessed EEG signal into a time-frequency two-dimensional image;

[0085] Specifically, this embodiment uses a wavelet transform method to convert the preprocessed EEG signal into a time-frequency two-dimensional image, which contains time domain information, frequency domain information and corresponding energy values ​​of the EEG signal.

[0086] Step 4: using a support vector machine classifier model to classify the time-frequency two-dimensional image;

[0087] In this embodiment, the purpose of classifying EEG signals is to identify the intention of motor imagery feature information and determine the category to which each motor imagery task belongs. This paper chooses to adopt the support vector machine classification method, uses grid search to automatically find the optimal parameters of the support vector machine, and finally classifies the motor imagery feature vector input to the classifier;

[0088] Specifically, Figure 4 As shown, using the support vector machine classifier model to classify the time-frequency two-dimensional image specifically includes:

[0089] Step 4.1: Build a support vector machine classifier;

[0090] Support vector machine is a binary classification model. Its basic idea is to find a plane in geometric space that can divide data of different categories and ensure the largest interval. For a linearly separable data set, there can be an infinite number of planes, but there is only one with the largest interval. Therefore, the solution of support vector machine is to increase the signal dimension.

[0091] Specifically, the support vector machine classifier is:

[0092]

[0093] Wherein, w is the characteristic matrix of the input signal, x is the input signal, and x i is the i-th eigenvector, y i For category, a i is the intermediate quantity, b is the coefficient, K(x i , x) is the kernel function of the classifier, and n represents the number of input signals;

[0094] Step 4.2: Design the structure of the classifier model;

[0095] In the classifier structure design of this embodiment, a plurality of classifiers are used to form a classifier group, each classifier is run separately to perform classification prediction, and then the final result is determined by voting based on the classification results of all classifiers;

[0096] Specifically, this embodiment uses 6 classifiers to form a classifier model based on the characteristics of EEG signals and the consideration of the operating equipment;

[0097] Step 4.3: Find the optimal parameters through grid search method;

[0098] The existence of multiple classifiers means that if the support vector machine parameters are determined manually, there is no way to ensure that the fixed parameter values ​​are the optimal solution for each classifier, which is prone to overfitting. To avoid this situation, this embodiment uses a grid search method to automatically find the optimal parameters for the support vector machine parameters;

[0099] In this embodiment, the optimal parameters are determined mainly by performing grid search on the penalty factor coefficient and the kernel function parameter coefficient. The specific method of finding the optimal parameters is common knowledge in the art and will not be described in detail here.

[0100] Step 4.4: inputting the time-frequency two-dimensional image into the support vector machine classifier model to obtain a classification result;

[0101] Specifically, the classification results include: left hand movement, right hand movement.

[0102] Example 2

[0103] like Figure 5 As shown, the present invention also includes a system for classifying EEG signals during rehabilitation training. The system adopts the method for classifying EEG signals of the brain-computer interface cap for rehabilitation training in Embodiment 1, including:

[0104] An EEG signal acquisition module, used to collect EEG signals through a brain-computer interface cap;

[0105] A preprocessing module, used to preprocess the EEG signal collected in step 1;

[0106] A signal conversion module converts the preprocessed EEG signal into a time-frequency two-dimensional image;

[0107] The classification module uses a support vector machine classifier model to classify the time-frequency two-dimensional image.

[0108] Example 3

[0109] This embodiment includes a computer-readable storage medium, on which a program is stored. The program is executed by a processor to implement the method for classifying EEG signals during rehabilitation training of the first embodiment.

[0110] It should be understood by those skilled in the art that the embodiments of this article can be provided as methods, devices (equipment), or computer program products. Therefore, this article can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, this article can adopt the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program codes. Computer storage media include volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules or other data), including but not limited to RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassette, magnetic tape, disk storage or other magnetic storage device, or any other medium that can be used to store desired information and can be accessed by a computer. In addition, it is well known to those of ordinary skill in the art that communication media generally contain computer-readable instructions, data structures, program modules or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0111] This document is described with reference to flowcharts and / or block diagrams of methods, apparatuses (devices) and computer program products according to the embodiments of this document. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0112] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0113] Although the preferred embodiments of this invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of this invention.

[0114] Obviously, those skilled in the art can make various changes and modifications to this article without departing from the spirit and scope of this article. Thus, if these modifications and variations of this article fall within the scope of the claims and their equivalents, then the intention of this article also includes these modifications and variations.

[0115] Although the preferred embodiments of this invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of this invention.

[0116] Obviously, those skilled in the art can make various changes and modifications to this article without departing from the spirit and scope of this article. Thus, if these modifications and variations of this article fall within the scope of the claims and their equivalents, then the intention of this article also includes these modifications and variations.

Claims

1. Classification method of EEG signals from brain-computer interface caps for rehabilitation training, It is characterized in that The following steps are involved: Step 1: Collect EEG signals through the brain-computer interface cap; Step 2: Preprocess the EEG signal collected in step 1; Step 3: Convert the preprocessed EEG signal into a time-frequency two-dimensional image; Step 4: Use a support vector machine classifier model to classify the time-frequency two-dimensional image.

2. The method for classifying EEG signals of a brain-computer interface cap for rehabilitation training according to claim 1, It is characterized in that In step 1, an automatic tightener is provided on the brain-computer interface cap, and the automatic tightener is used to bring the electrodes of the brain-computer interface cap into close contact with human skin.

3. The method for classifying EEG signals of a brain-computer interface cap for rehabilitation training according to claim 2, It is characterized in that The automatic tightener includes a pressure sensor, a motor, and a microcontroller. The motor is arranged on the rear edge of the brain-computer interface cap and is connected to the microcontroller. An automatic tightening wire is wound on the motor output shaft. The free end of the automatic tightening wire is connected to the wedge block through the automatic tightener wire. The wedge block is arranged in the transverse channel of the mounting block. The mounting block is arranged on the inner wall of the brain-computer interface cap, and the mounting block is also provided with a vertical channel, which is connected to the transverse channel, and an electrode push rod is arranged in the vertical channel. The top of the electrode push rod is arranged as an inclined surface and contacts the inclined surface of the wedge block, and an electrode is arranged at the bottom thereof. The pressure sensor is arranged on the motor push rod.

4. The method for classifying EEG signals of a brain-computer interface cap for rehabilitation training according to claim 3, It is characterized in that The wedge block is an I-shaped structure, the vertical plates at both ends of which are located outside the mounting block, and a compression spring is sleeved on the wedge block, which is located outside the mounting block and connected to the side wall of the mounting block.

5. The method for classifying EEG signals of a brain-computer interface cap for rehabilitation training according to claim 1, It is characterized in that The preprocessing of the EEG signal collected in step 1 specifically includes: Step 2.1: Filter the EEG signal; Step 2.2: Perform baseline correction on the EEG signal processed in step 2.1; Step 2.3: Fill the data of the damaged channel in the EEG signal processed in step 2.2; Step 2.4: Use principal component analysis to further filter the EEG signal processed in step 2.

3.

6. The method for classifying EEG signals of a brain-computer interface cap for rehabilitation training according to claim 1, It is characterized in that The support vector machine classifier model is used to classify the time-frequency two-dimensional image, specifically including: Step 4.1: Build a support vector machine classifier; Step 4.2: Design the structure of the support vector machine classifier model; Step 4.3: Find the optimal parameters through grid search method; Step 4.4: Input the time-frequency two-dimensional image into the support vector machine classifier model to obtain the classification result.

7. The method for classifying EEG signals of a brain-computer interface cap for rehabilitation training according to claim 6, It is characterized in that The support vector machine classifier is: In the formula, w is the characteristic matrix of the input signal, x is the input signal, and x i is the i-th eigenvector, y i For category, a i is the intermediate quantity, b is the coefficient, K(x i , x) is the kernel function of the classifier, and n represents the number of input signals.

8. The method for classifying EEG signals of a brain-computer interface cap for rehabilitation training according to claim 5, It is characterized in that The optimal parameters are determined by grid searching the penalty factor coefficient and the kernel function parameter coefficient.

9. Brain-computer interface cap EEG signal classification system for rehabilitation training, It is characterized in that The system includes: An EEG signal acquisition module, used to acquire EEG signals through a brain-computer interface cap; A preprocessing module, used for preprocessing the EEG signals collected by the EEG signal collection module; A signal conversion module converts the preprocessed EEG signals into a time-frequency two-dimensional image; The classification module calls the support vector machine classifier model to classify the time-frequency two-dimensional image, and the support vector machine classifier model is stored in the storage module.

10. A computer-readable storage medium having a program stored thereon, wherein the program is executed by a processor to implement the method for classifying EEG signals of a brain-computer interface cap for rehabilitation training as described in any one of claims 1 to 8.

Images