Method, apparatus, and computer program for creating standardized brainwave image for training artificial intelligence model

1 . A method of creating a standardized brainwave image for training an artificial intelligence model executed by a computing device, the method comprising:

collecting a plurality of brainwave signals of a user in a plurality of frequency bands through a plurality of brainwave measurement channels, wherein the plurality of brainwave measurement channels are configured to be attached to a plurality of locations of a head of the user;

processing the plurality of collected brainwave signals;

creating a brainwave image using the plurality of processed brainwave signals; and

performing training of an artificial intelligence (AI) model based on the brainwave image, and using the AI model to generate output data indicative of a state information of another user based on input of another user,

wherein processing the plurality of collected brainwave signals includes:

calculating index values for the plurality of collected brainwave signals;

preprocessing the plurality of brainwave signals from which the index values are calculated, and filtering, among the plurality of brainwave signals from which the index values are calculated, a brainwave signal in a predetermined frequency band; and

determining, based on the calculated index values, pixel values corresponding to the plurality of preprocessed brainwave signals respectively,

wherein creating the brainwave image includes creating the brainwave image by using the determined pixel values,

wherein creating the brainwave image by using the determined pixel values includes creating the brainwave image by arranging, in a preset template, the pixel values determined for the plurality of preprocessed brainwave signals respectively,

wherein, in the preset template, pixel values according to a change in frequency are arranged on a first axis, and pixel values according to a change in position of a region of interest of a brain of the user are arranged on a second axis perpendicular to the first axis,

wherein a preset reference frequency value is set as a central axis of the first axis,

wherein pixel values corresponding to brainwave signals collected from a left region of the brain are arranged in a left region with respect to the central axis, and pixel values corresponding to brainwave signals collected from a right region of the brain are arranged in a right region with respect to the central axis,

wherein creating the brainwave image by arranging the determined pixel values in the preset template includes:

dividing the first axis, wherein the dividing the first axis comprises:

generating a plurality of unit sections by dividing, based on frequency bands, a left side of the first axis with respect to the central axis, and further dividing, based on the frequency bands, a right side of the first axis with respect to the central axis, the frequency bands corresponding to types of the brainwave signals,

wherein the plurality of unit sections have a same length regardless of a range of the frequency bands corresponding to the types of the brainwave signals, and

wherein dividing the first axis includes:

based on a request to extend a length of a first unit section among the plurality of unit sections, extending the length of the first unit section, and shortening lengths of the remaining unit sections by the extended length of the first unit section, and

based on a request to shorten a length of a second unit section among the plurality of unit sections, shortening the length of the second unit section, and extending the lengths of the remaining unit sections by the shortened length of the second unit section.

2 . The method of claim 1 , wherein processing the plurality of collected brainwave signals includes:

calculating index values for the plurality of collected brainwave signals;

preprocessing the plurality of brainwave signals from which the index values are calculated, and filtering, among the plurality of brainwave signals from which the index values are calculated, a brainwave signal in a predetermined frequency band; and

determining, based on the calculated index values, pixel values corresponding to the plurality of preprocessed brainwave signals respectively.

3 . The method of claim 2 , wherein calculating the index values includes:

calculating the index values using at least one of absolute power, relative power, a standard value (Z-score), complexity, and entropy of the plurality of preprocessed brainwave signals, respectively.

4 . The method of claim 1 , wherein creating the brainwave image by arranging the determined pixel values in the preset template includes:

generating a plurality of unit sections by dividing the first axis based on frequency bands, the frequency bands corresponding to types of the brainwave signals, and

determining a length of each of the plurality of unit sections based on importance of each of the frequency bands, the importance being input by a user.

5 . The method of claim 1 , wherein creating the brainwave image by arranging the determined pixel values in the preset template includes:

generating a pixel value matrix using the plurality of pixel values determined to correspond to the plurality of preprocessed brainwave signals respectively, the generated pixel value matrix having (i) pixel values being arranged in a row direction according to a change in frequency of the plurality of preprocessed brainwave signals and (ii) pixel values being arranged in a column direction according to a change in position of a region of interest of the plurality of preprocessed brainwave signals;

converting the generated pixel value matrix into a square matrix; and

arranging the square matrix in the preset template to generate the standardized brainwave image.

6 . The method of claim 1 , wherein creating the brainwave image by arranging the determined pixel values in the preset template includes:

arranging the plurality of pixel values, determined to correspond to the plurality of preprocessed brainwave signals, in a region that corresponds to a frequency and a position of a point of interest of each of the plurality of preprocessed brainwave signals on the preset template, and performing image smoothing on pixel values arranged adjacent to the plurality of pixel values; and

creating the standardized brainwave image by performing image resizing that converts the image-smoothed brainwave image to a square having four sides of equal length.

7 . The method of claim 6 , further comprising:

generating training data for training the artificial intelligence model by using the created standardized brainwave image,

wherein the generated training data includes one standardized brainwave image or a plurality of standardized brainwave images combined in a square having four side of equal length.

8 . An apparatus for creating a standardized brainwave image for training an artificial intelligence model, the apparatus comprising:

a processor,

a memory; and

a computer program loaded into the memory and executed by the processor to cause the apparatus to perform operations comprising:

collecting a plurality of brainwave signals of a user in a plurality of frequency bands through a plurality of brainwave measurement channels, wherein the plurality of brainwave measurement channels are configured to be attached to a plurality of locations of a head of the user;

processing the plurality of collected brainwave signals;

creating a brainwave image using the plurality of processed brainwave signals; and

performing training of an artificial intelligence (AI) model based on the brainwave image, and using the AI model to generate output data indicative of a state information of another user based on input of another user,

wherein processing the plurality of collected brainwave signals includes:

calculating index values for the plurality of collected brainwave signals,

preprocessing the plurality of brainwave signals from which the index values are calculated, and filtering, among the plurality of brainwave signals from which the index values are calculated, a brainwave signal in a predetermined frequency band, and

determining, based on the calculated index values, pixel values corresponding to the plurality of preprocessed brainwave signals respectively,

wherein creating the brainwave image includes creating the brainwave image by using the determined pixel values,

wherein creating the brainwave image by using the determined pixel values includes creating the brainwave image by arranging, in a preset template, the pixel values determined for the plurality of preprocessed brainwave signals respectively, and

wherein, in the preset template, pixel values according to a change in frequency are arranged on a first axis, and pixel values according to a change in position of a region of interest of a brain of the user are arranged on a second axis perpendicular to the first axis,

wherein a preset reference frequency value is set as a central axis of the first axis,

wherein pixel values corresponding to brainwave signals collected from a left region of the brain are arranged in a left region with respect to the central axis, and pixel values corresponding to brainwave signals collected from a right region of the brain are arranged in a right region with respect to the central axis,

wherein creating the brainwave image by arranging the determined pixel values in the preset template includes:

dividing the first axis, wherein the dividing the first axis comprises:

generating a plurality of unit sections by dividing, based on frequency bands, a left side of the first axis with respect to the central axis, and further dividing, based on the frequency bands, a right side of the first axis with respect to the central axis, the frequency bands corresponding to types of the brainwave signals,

wherein the plurality of unit sections have a same length regardless of a range of the frequency bands corresponding to the types of the brainwave signals, and

wherein dividing the first axis includes:

based on a request to extend a length of a first unit section among the plurality of unit sections, extending the length of the first unit section, and shortening lengths of the remaining unit sections by the extended length of the first unit section, and

based on a request to shorten a length of a second unit section among the plurality of unit sections, shortening the length of the second unit section, and extending the lengths of the remaining unit sections by the shortened length of the second unit section.

9 . A non-transitory computer-readable medium having stored therein a computer program for causing a computing device to execute a process of creating a standardized brainwave image for training an artificial intelligence model, the process comprising:

collecting a plurality of brainwave signals of a user in a plurality of frequency bands through a plurality of brainwave measurement channels, wherein the plurality of brainwave measurement channels are configured to be attached to a plurality of locations of a head of the user;

processing the plurality of collected brainwave signals;

creating a brainwave image using the plurality of processed brainwave signals; and

performing training of an artificial intelligence (AI) model based on the brainwave image, and using the AI model to generate output data indicative of a state information of another user based on input of another user,

wherein processing the plurality of collected brainwave signals includes:

calculating index values for the plurality of collected brainwave signals,

preprocessing the plurality of brainwave signals from which the index values are calculated, and filtering, among the plurality of brainwave signals from which the index values are calculated, a brainwave signal in a predetermined frequency band, and

determining, based on the calculated index values, pixel values corresponding to the plurality of preprocessed brainwave signals respectively,

wherein creating the brainwave image includes creating the brainwave image by using the determined pixel values,

wherein creating the brainwave image by using the determined pixel values includes creating the brainwave image by arranging, in a preset template, the pixel values determined for the plurality of preprocessed brainwave signals respectively, and

wherein, in the preset template, pixel values according to a change in frequency are arranged on a first axis, and pixel values according to a change in position of a region of interest of a brain of the user are arranged on a second axis perpendicular to the first axis,

wherein a preset reference frequency value is set as a central axis of the first axis,

wherein pixel values corresponding to brainwave signals collected from a left region of the brain are arranged in a left region with respect to the central axis, and pixel values corresponding to brainwave signals collected from a right region of the brain are arranged in a right region with respect to the central axis,

wherein creating the brainwave image by arranging the determined pixel values in the preset template includes:

dividing the first axis, wherein the dividing the first axis comprises:

generating a plurality of unit sections by dividing, based on frequency bands, a left side of the first axis with respect to the central axis, and further dividing, based on the frequency bands, a right side of the first axis with respect to the central axis, the frequency bands corresponding to types of the brainwave signals,

wherein the plurality of unit sections have a same length regardless of a range of the frequency bands corresponding to the types of the brainwave signals, and

wherein dividing the first axis includes:

based on a request to extend a length of a first unit section among the plurality of unit sections, extending the length of the first unit section, and shortening lengths of the remaining unit sections by the extended length of the first unit section, and

based on a request to shorten a length of a second unit section among the plurality of unit sections, shortening the length of the second unit section, and extending the lengths of the remaining unit sections by the shortened length of the second unit section.