KR101879408B1 - Apparatus and method for determining optimal frequency area in brain computer interfacing - Google Patents

Abstract

The present invention discloses a brain computer interface (BCI) device for recognizing a motion intention of a subject. The BCI apparatus includes a memory for storing a motion intention program; And a processor for executing the program, wherein the processor acquires a brain signal measured from a subject as the program is executed, and then calculates a brain signal feature value corresponding to each of a plurality of predetermined frequency ranges, The classification performance of each frequency region is ranked based on the brain signal feature value and at least one lower frequency region having a classification performance lower than the threshold value is shifted by a predetermined value shift results.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and apparatus for determining a frequency range optimized for a subject in a BCI technique for recognizing motion intent,

The present invention relates to a method and apparatus for determining a frequency domain optimized for a subject in a BCI (Brain-Computer Interface) technique for recognizing a motion intention.

The BCI technology for motion intention recognition can be a computer or a machine control by measuring a brain signal that is expressed as a user imagines the movement of the body or intends to move the body.

In general, BCI technology measures EEG in various frequency bands that are expressed by the user's motion intention. Specifically, when the user intends (or imagines) motion, the primary motor cortex of the brain is activated and the mu-band of the frequency region of about 8 to 12 Hz reacts. Therefore, the BCI technology recognizes the user's motion intention using the change of mu-band which is a specific frequency range.

However, the frequency region of the EEG that is reacted by the user's motion intention may be different for each user. Therefore, there is an increasing need for a method and device for recognizing a user's motion intention based on characteristics of brain waves for each user.

On the other hand, Korean Patent Registration No. 10-1518575 (Registration Date: 2015.04.30, entitled "User Intention Cognitive Analysis Method for BCI") is known as a conventional technique for recognizing a user's motion intention. This is because it can not extract the optimized frequency domain for each subject in that it starts extracting the brain signal of the mu band band by using the digital notch filter and the band pass filter There is a limit.

It is an object of the present invention to provide a method and an apparatus for determining an optimal frequency domain for each subject to improve the accuracy of a motion intention recognition technique.

As a technical means for achieving the above-mentioned technical object, a first aspect of the present invention is a memory device, comprising: a memory for storing the motion intention recognizing program; And a processor for executing the program. At this time, the processor acquires the brain signal measured from the subject according to the execution of the program, calculates a brain signal characteristic value corresponding to each of a plurality of preset frequency ranges, The classification performance of the frequency domain is ranked and at least one lower frequency region having the classification performance equal to or less than the threshold value is replaced with a result value obtained by shifting the upper frequency region having the highest classification performance by a preset value .

The processor may filter the obtained brain signals to the respective frequency regions and calculate the frequency power of the filtered brain signals.

The processor is configured to generate a brain signal corresponding to each of a plurality of frequency regions in which at least one lower frequency region is replaced until the difference between the classification performance of at least one lower frequency region and the classification performance of the upper frequency region falls within a critical range, The feature value can be repeatedly calculated.

Wherein the processor is configured to generate a brain signal characteristic value corresponding to the entire plurality of frequency regions in which at least one lower frequency region is replaced if the difference between the classification performance of at least one lower frequency region and the classification performance of the upper frequency region is within a critical range Generates a classifier based on the brain signal feature value, and recognizes the motion intention from the brain signal measured from the subject using the classifier.

The processor compares the minimum value and the minimum value of the lower frequency range with the minimum value and the highest value in the upper frequency range and shifts the upper frequency range in the positive direction or the negative direction according to the result of the comparison. have. If the maximum value of the lower frequency domain is smaller than the minimum value of the upper frequency domain, the processor replaces the lower frequency domain with the resultant value shifted in the negative (-) direction, If it is larger than the maximum value of the frequency domain, the lower frequency domain can be replaced with the shifted result in the positive (+) direction.

The processor includes a cross-validation method using a classifier learned with a brain signal feature value, mutual information of brain signal feature values of each class included in the classifier, The classification performance of each frequency region can be calculated based on at least one of the Fisher's score according to the distance of the value.

According to a second aspect of the present invention, there is provided a method for generating a brain signal, comprising the steps of: (a) acquiring a brain signal measured from a subject and calculating a brain signal feature value corresponding to each of a plurality of predetermined frequency ranges; (b) ranking the classification performance of each frequency domain based on the brain signal feature value; And (c) replacing a lower frequency region having a classification performance lower than a threshold value with a result value obtained by shifting an upper frequency region having the highest classification performance by a preset value. ≪ / RTI >

A third aspect of the present invention provides a computer-readable recording medium recording a program for performing the method of the second aspect on a computer.

Conventional motion intention recognition technology has a limitation in that only a specific frequency band is considered regardless of the characteristics of each subject. In order to solve such conventional problems, the present invention can provide a more optimized motion intention recognition technique for a subject by searching not only a specific frequency band but a frequency band suitable for the subject.

1 is a block diagram illustrating a configuration of a BCI apparatus according to an embodiment of the present invention.
FIG. 2 is a flowchart illustrating a method of determining a frequency region optimized for a subject by the processor of FIG. 1 according to an embodiment of the present invention.
FIG. 3 is an example of a process of determining a frequency region optimized for a subject according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings, which will be readily apparent to those skilled in the art. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In order to clearly illustrate the present invention, parts not related to the description are omitted, and similar parts are denoted by like reference characters throughout the specification.

Throughout the specification, when a part is referred to as being "connected" to another part, it includes not only "directly connected" but also "electrically connected" with another part in between . Also, when an element is referred to as "comprising ", it means that it can include other elements as well, without departing from the other elements unless specifically stated otherwise.

In this specification, the term " part " includes a unit realized by hardware, a unit realized by software, and a unit realized by using both. Further, one unit may be implemented using two or more hardware, or two or more units may be implemented by one hardware. On the other hand, 'to' is not limited to software or hardware, 'to' may be configured to be an addressable storage medium, and may be configured to play one or more processors. Thus, by way of example, 'parts' may refer to components such as software components, object-oriented software components, class components and task components, and processes, functions, , Subroutines, segments of program code, drivers, firmware, microcode, circuitry, data, databases, data structures, tables, arrays, and variables. The functions provided in the components and components may be further combined with a smaller number of components and components or further components and components. In addition, the components and components may be implemented to play back one or more CPUs in a device or a secure multimedia card.

The "BCI (Brain Computer Interface) device" mentioned below may be implemented as a computer or a portable terminal. Here, the computer includes, for example, a notebook, a desktop, a laptop, a medical device, and the like, and the portable terminal is, for example, a portable and mobility-guaranteed device such as a smart phone, a tablet PC, And may include all sorts of handheld based devices. Further, the BCI apparatus may be implemented as a wearable apparatus in a form in which a user (or a subject) can be worn. In addition, the BCI device can access a server or other terminal through a network. Herein, the term "network" refers to a wired network such as a local area network (LAN), a wide area network (WAN) or a value added network (VAN), a mobile radio communication network, And may be implemented in all kinds of wireless networks, such as communication networks.

Further, the "motor imagery" referred to below may be referred to as an operational image, a moving image, a dynamic image, and the like.

Hereinafter, one embodiment of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a diagram showing a configuration of a BCI apparatus 100 according to an embodiment of the present invention. Referring to FIG. 1, a BCI apparatus 100 according to an embodiment of the present invention includes a brain signal measuring unit 110, a processor 120, and a memory 130.

The brain signal measuring unit 110 may be embodied as at least one or more components for measuring brain signals (i.e., brain waves) of the subject. For example, the brain signal measuring unit 110 may be an EEG (Electroencephalography) apparatus that non-invasively detects the brain waves of the subject.

The brain signal measuring unit 110 can measure the brain signal expressed by the motion intention of the subject from the scalp of the subject. The brain signal measuring unit 110 may perform a signal processing operation on the measured brain signal, and then provide the signal processed brain signal to the processor 120.

The memory 130 stores a control program for controlling the BCI apparatus 100. In addition, the memory 130 stores a motion intention program for determining a motion intention based on a brain signal measured from a subject. More specifically, the program may include one or more instructions to control the processor 120 to determine a frequency domain optimized for the subject based on a brain signal. The program can be read out and executed by the processor 120. [

Meanwhile, the memory 130 may collectively refer to an inactive storage device that maintains stored information even when power is not supplied, and a volatile storage device that requires power to maintain stored information.

The processor 120 may be implemented with at least one processing unit (CPU, micro-processor, DSP, etc.), random access memory (RAM), read-only memory (ROM) The program can be read into the RAM and executed via at least one processing unit. In the meantime, the term 'processor' may be interpreted in the same sense as the terms 'controller', 'controller', 'arithmetic unit', and the like.

Hereinafter, a method for determining the frequency region optimized for the subject by executing the program stored in the memory 130 by the processor 120 will be described with reference to FIG. Referring to FIG. 2, the brain signal measuring unit 110 measures the brain signal of the subject (S110). The measured brain signal is provided to the processor 120 after a preprocessing process.

After obtaining the brain signal, the processor 120 extracts a feature value of the brain signal corresponding to each of a plurality of predetermined frequency ranges (S120). At this time, a plurality of preset frequency ranges may be obtained by dividing experimentally measured frequency bands at regular intervals with respect to a standard standard. For example, a plurality of frequency regions may be [1-5 Hz, 6-10 Hz, ..., 45-50 Hz] in which 1 to 50 Hz regions are separated by 5 Hz intervals.

The processor 120 filters the brain signal into the plurality of frequency regions. The processor 120 then calculates the frequency power of the filtered brain signal. That is, the feature value of the brain signal corresponding to each frequency domain may be the power value of the brain signal filtered in each frequency domain.

Next, the processor 120 ranks the classification performance of each frequency domain based on the feature value (S130). Here, the classification performance of each frequency domain may mean classification accuracy in which the classifier classifies the motion intention. Specifically, the processor 120 can learn the classifier using the feature values of each frequency domain and calculate the classification accuracy using the cross-validation. However, the present invention is not limited thereto, and the processor 120 may calculate Mutual information of the feature values of the classes included in the classifier, or calculate Fisher's score according to the distance of each feature value of the classes, The classification accuracy can be calculated.

Next, the processor 120 replaces at least one lower frequency region having a classification performance lower than the threshold with a result value obtained by shifting the upper frequency region having the highest classification performance by a predetermined value (S140). The predetermined value may be, for example, about one-tenth the value of each frequency domain size. That is, when each frequency domain has a size of 5 Hz, the preset value may be 0.5 Hz. Therefore, when the frequency region having the lowest classification performance is [1-5 Hz] and the frequency region having the highest classification performance is [11-15 Hz], the [1-5 Hz] region is replaced with [11.5-15.5 Hz] . However, the present invention is not limited thereto, and the processor 120 may shift the upper frequency domain to any value.

In addition, in the above description, the processor 120 shifts the upper frequency region in the positive direction to replace the lower frequency region, but is not limited thereto. For example, the processor 120 may shift the upper frequency domain in the negative direction. The processor 120 compares the minimum value and the minimum value of the lower frequency range with the minimum value and the minimum value of the upper frequency range and shifts the upper frequency range in the positive direction or the negative direction according to the comparison result, can do. Specifically, when the maximum value of the lower frequency region is smaller than the minimum value of the upper frequency region, the processor 120 replaces the lower frequency region with a value shifted in the negative direction, Is higher than the maximum value of the upper frequency domain, the upper frequency domain can be replaced with a value shifted in the positive direction to replace the lower frequency domain. For example, in this case, the [1-5 Hz] region may be replaced with a [10.5-14.5 Hz] region shifted in the negative direction by [11-15 Hz].

Meanwhile, the processor 120 repeatedly performs the above-described steps S120 and S140 until the difference between the classification performance of the upper frequency domain and the classification performance of the lower frequency domain falls within the critical range, thereby repeatedly updating the plurality of frequency domains. Through this, the processor 120 can determine the frequency range optimized for the subject. That is, if the difference between the classification performance of the upper frequency domain and the classification performance of the lower frequency domain is within the critical range, the processor 120 calculates the brain signal characteristic value corresponding to the entire updated frequency domain. Thereafter, the processor 120 may generate a classifier using only the feature values corresponding to the entire frequency domain, thereby improving the motion intention / performance of the subject. In addition, the frequency domain optimized for each subject may be stored in the memory 130 and used repeatedly.

1 and 2, the BCI apparatus 100 includes the brain signal measuring unit. However, according to the embodiment, the BCI apparatus 100 may be configured as a separate apparatus from the brain signal measuring unit . In this case, the BCI apparatus 100 can acquire the brain signal measured by the brain signal measuring unit through a wired / wireless network (e.g., a wired cable, a wireless communication, and the like).

FIG. 3 is an example showing a process of determining a frequency region optimized for a subject by the BCI apparatus 100. FIG.

Referring to FIG. 3, the processor 120 optimizes the first through tenth frequency regions in which 1 Hz to 50 Hz regions are divided at 5 Hz intervals into regions suitable for the subject. Specifically, the processor 120 determines a lower frequency region having a classification performance lower than the threshold value and an upper frequency region having the highest classification performance based on the classification characteristics calculated based on the brain signal characteristic values corresponding to the respective frequency regions, . For example, as shown in FIG. 3 (a), the upper frequency range may be [11-15 Hz] and the lower frequency range may be [1-5 Hz] and [46-50 Hz]. Accordingly, the processor 120 replaces the lower frequency region [1-5 Hz] with the [10.5-14.5 Hz] region and replaces the other lower frequency region [46-50 Hz] with the [11.5-15.5 Hz] region, The area can be updated.

Then, the processor 120 repeats the above process to extract the upper frequency region [11-15 Hz] and the lower frequency region [6-10 Hz] and [41-45 Hz] as shown in FIG. 3 (b) can do. Accordingly, the processor 120 can repeatedly replace the lower frequency region [6-10 Hz] with the [10.5-14.5 Hz] region and replace [41-46 Hz] with the [11.5-15.5 Hz] region.

Meanwhile, according to an embodiment, the BCI apparatus 100 can adjust a magnitude value that shifts an upper frequency region every time a plurality of frequency regions are updated. For example, the processor 120 may shift the upper frequency domain to a different value according to the number of times of updating the plurality of frequency domains. For example, referring to FIG. 3, the processor 120 may replace the lower frequency region with a value obtained by shifting the upper frequency region by '0.5 + x * i (x is a preset rational number and i is the number of updates)' . That is, the processor 120 shifts the upper frequency region by a value of 0.5 in FIG. 3 (a), shifts by 0.5 + x * 1 in FIG. 3 (b) 0.5 + x * 2 '.

As described above, the BCI apparatus 100 according to the disclosed embodiment can more accurately recognize the motion intention of the subject by determining the frequency region optimized for the subject, and generating the corresponding classifier.

One embodiment of the present invention may also be embodied in the form of a recording medium including instructions executable by a computer, such as program modules, being executed by a computer. Computer readable media can be any available media that can be accessed by a computer and includes both volatile and nonvolatile media, removable and non-removable media. The computer-readable medium may also include computer storage media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data.

While the methods and systems of the present invention have been described in connection with specific embodiments, some or all of those elements or operations may be implemented using a computer system having a general purpose hardware architecture.

It will be understood by those skilled in the art that the foregoing description of the present invention is for illustrative purposes only and that those of ordinary skill in the art can readily understand that various changes and modifications may be made without departing from the spirit or essential characteristics of the present invention. will be. It is therefore to be understood that the above-described embodiments are illustrative in all aspects and not restrictive. For example, each component described as a single entity may be distributed and implemented, and components described as being distributed may also be implemented in a combined form.

The scope of the present invention is defined by the appended claims rather than the detailed description and all changes or modifications derived from the meaning and scope of the claims and their equivalents are to be construed as being included within the scope of the present invention do.

100: BCI device
110: brain signal measuring unit
120: Processor
130: memory

Claims (13)

A brain computer interface (BCI) device for recognizing a motion intention of a subject,
A memory in which the motion intention aware program is stored; And
And a processor for executing the program,
The processor, as the program is executed,
A brain signal characteristic value corresponding to each of a plurality of predetermined frequency ranges is calculated after acquiring a brain signal measured from a subject,
Ranking the classification performance of each frequency region based on the brain signal feature value,
Wherein at least one lower frequency region having a classification performance equal to or lower than a threshold value is replaced with a result value obtained by shifting an upper frequency region having the highest classification performance by a preset value. The method according to claim 1,
The processor
Filtering the acquired brain signals to the respective frequency regions, and calculating a frequency power of the filtered brain signals. The method according to claim 1,
The processor
Wherein the at least one lower frequency domain has a brain signal characteristic corresponding to each of the plurality of frequency regions replaced by the at least one lower frequency domain until the difference between the classification performance of the at least one lower frequency domain and the classification performance of the upper frequency domain is within a critical range, Value of the BCI device. The method of claim 3,
The processor
If a difference between the classification performance of the at least one lower frequency region and the classification performance of the upper frequency region is within a critical range, a brain signal characteristic value corresponding to the entire plurality of frequency regions in which the at least one lower frequency region is replaced is calculated Generates a classifier based on the brain signal feature value, and recognizes the motion intent from the brain signal measured from the subject using the classifier. The method according to claim 1,
The processor
The minimum value and the maximum value of the lower frequency range are compared with each other, and the upper frequency range is shifted in the positive (+) direction or the negative (-) direction according to the comparison result BCI device. 6. The method of claim 5,
The processor
If the maximum value of the lower frequency region is smaller than the minimum value of the upper frequency region, the lower frequency region is replaced with the shifted result value in the direction of the upper frequency region,
And if the minimum value of the lower frequency domain is greater than the maximum value of the upper frequency domain, the lower frequency domain is replaced with the shifted result value in the positive direction. The method according to claim 1,
The processor
A cross-validation using the classifier learned by the brain signal feature value, a mutual information of brain signal feature values of each class included in the classifier, and a brain signal feature value of each of the classes (Fisher ' s ' score) according to the distance between the BCI device and the BCI device. A method of operating a brain computer interface (BCI) apparatus for recognizing a motion intention of a subject,
(a) obtaining a brain signal measured from a subject, and calculating a brain signal characteristic value corresponding to each of a plurality of preset frequency ranges;
(b) ranking the classification performance of each frequency domain based on the brain signal feature value; And
(c) replacing a lower frequency region having a classification performance lower than a threshold value with a result value obtained by shifting an upper frequency region having the highest classification performance by a preset value. 9. The method of claim 8,
The step (a)
(a-1) filtering the acquired brain signal to the respective frequency regions; And
(a-2) calculating a frequency power of the filtered brain signal. 9. The method of claim 8,
The operating method
And repeating the steps (a) to (c) until the difference between the classification performance of the at least one lower frequency region and the classification performance of the upper frequency region falls within a threshold range. 11. The method of claim 10,
The operating method
(d) if the difference between the classification performance of the at least one lower frequency domain and the classification performance of the upper frequency domain is within a critical range, the at least one lower frequency domain has a brain signal characteristic Calculating a value;
(e) generating a classifier based on the brain signal feature value; And
(f) recognizing movement intent from a brain signal measured from the subject using the classifier. 9. The method of claim 8,
The step (c)
(c-1) comparing a minimum value and a minimum value of the lower frequency range with a minimum value and a minimum value of the upper frequency range; And
(c-2) shifting the upper frequency region in a positive direction or a negative direction according to a result of the comparison; ≪ / RTI > A computer-readable recording medium recording a program for performing the method according to any one of claims 8 to 12 on a computer.

Images