CN105045898B - A kind of dynamically personalized song recommendations system and method based on brain wave earphone - Google Patents

A kind of dynamically personalized song recommendations system and method based on brain wave earphone Download PDF

Info

Publication number
CN105045898B
CN105045898B CN201510467034.6A CN201510467034A CN105045898B CN 105045898 B CN105045898 B CN 105045898B CN 201510467034 A CN201510467034 A CN 201510467034A CN 105045898 B CN105045898 B CN 105045898B
Authority
CN
China
Prior art keywords
song
weights
step
brain
system
Prior art date
Application number
CN201510467034.6A
Other languages
Chinese (zh)
Other versions
CN105045898A (en
Inventor
曾洪
石春凤
杨骏逸
端豪
郑亚君
张艺璇
宋爱国
Original Assignee
东南大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 东南大学 filed Critical 东南大学
Priority to CN201510467034.6A priority Critical patent/CN105045898B/en
Publication of CN105045898A publication Critical patent/CN105045898A/en
Application granted granted Critical
Publication of CN105045898B publication Critical patent/CN105045898B/en

Links

Abstract

The dynamically personalized song recommendations system and method based on brain wave earphone that the invention discloses a kind of, system includes brain wave earphone system, user terminal and server, the brain feedback data that the brain wave earphone system built-in algorithm obtains is transmitted to user terminal, the brain feedback data received is transmitted to server by user terminal, after server receives brain feedback data, brain feedback data is carried out to assign weights processing by the song recommendations algorithm improved, it is resequenced to song according to updated weights, is formed and song patrilineal line of descent with only one son in each generation is recommended to transport to user terminal.The music that can be adapted for the various states such as work, study and sleep using the method for present system is recommended, and a more hommization and targetedly man-machine system is created that for user.

Description

A kind of dynamically personalized song recommendations system and method based on brain wave earphone

Technical field

The present invention relates to a kind of dynamically personalized song recommendations system and method, and MindWave is based on more particularly to one kind The individualized music commending system and its method of brain wave earphone.

Background technology

Custom finder by observing people is accustomed to reading, learn or working while listening music.It closes Suitable music helps to improve the focus of people, improves work, learning efficiency;On the contrary, song that is excessively fierce or excessively releiving Song can influence the thinking ability of people, reduce focus, be unfavorable for the progress of work or study.If music playing system is not Effectively recommended for personalized preference, it may be because of having played the song not liked or not met the sound of mood at that time It is happy, and cause to interrupt task at hand to adjust music, and this easily makes people can not be wholwe-hearted when reading, and generates study work Make the negative effects such as inefficiency, therefore it is a great application value that dynamically personalized music recommendation, which how is effectively performed, And the problem of foreground.

It is the recommendation based on keyword or fixed scene pattern mostly, for user in current music commending system There is no specific aim.It can not recommend song according to personal emotional change.How the emotional change of music to be selected according to hearer It selects and recommends song automatically, be the new way of music commending system.

There are some music commending systems or method based on emotion in the prior art.Bibliography:[1] it is based on brain-machine interaction Music moodization recommend method CN103412646A (Wang Wei Yuan sea cloud summer chess Gao Jia) 8/7.2013, be with mood be index, It is different in the frequency in different songs by EEG signals, it is identified with trained brain electricity Emotion identification model, Judge mood classification, corresponding song emotion characteristic search combination is carried out to song, and then classification is handled, and realizes the mood to song Change and recommend, however this method is only capable of realizing the recommendation under several discrete emotional states, it is clear that can not include all in reality Emotional state.[2] a kind of method for searching music CN101226526A (Zhang Wenqi Cheng Wei people's models based on music clip information inquiry Enlightening) 23July2008. [2] ChangshengXu.Automatic music classifica tion and summarization.[J]Speech and Audio Processing,2005,V13(3):441-450.[3]Mu Li& Bao-LiangLu.Emotion Classification Based on Gamma-band EEG[J],31st Annual International Conference of the IEEE EMBS,Minineapolis,Minnesota,UDA September206,2009, p132301326), studied the acoustic feature of music, it is intended to and emotion type is found out, into And according to emotion type classification.But such method judges relatively to fix for emotion, can not be changed according to user emotion state And adaptively dynamic adjustment song is single.And using user to the EEG signals of song variation automatic decision song if appropriate for push away Recommend method not yet.User being heard to, the change detection of brain wave when song comes out, and carries out assignment processing, so that it may judge song It is bent whether to be suitble to by user, in a manner of being indexed by mood, keep commending system more acurrate.With the development of brain wave earphone, It reads user's eeg data and judges that current state has become possibility.

Invention content

Goal of the invention:In order to overcome the deficiencies in the prior art, the present invention to provide a kind of dynamic based on brain wave earphone State personalization song recommendations system and method, the music that can be adapted for the various states such as work, study and sleep are recommended, It is created that a more hommization and targetedly man-machine system for user.

Technical solution:To achieve the above object, the technical solution adopted by the present invention is:

A kind of dynamically personalized song recommendations system based on brain wave earphone, including brain wave earphone system, user terminal and Server, the brain feedback data that the brain wave earphone system built-in algorithm obtains are transmitted to user terminal, and user terminal will connect The brain feedback data received is transmitted to server, after server receives brain feedback data, passes through the song recommendations improved Algorithm carries out brain feedback data to assign weights processing, is resequenced to song according to updated weights, is formed and recommend song single It is transmitted to user terminal.The brain feedback data includes focus index x and allowance index y.

A method of the dynamically personalized song recommendations system based on brain wave earphone is as follows:

The first step:The user terminal reads the brain feedback data that brain wave earphone system-computed obtains by bluetooth, point Not Wei focus index x and allowance index y, and be uploaded to server;

Second step:The brain feedback data focus index X and allowance index Y that the server receives the first step into Row normalized;

Third walks:Point is calculated away from according to different situations progress right value update by the result after calculating second step normalization Processing;

4th step:Calculated new weights are walked according to third to be ranked up song, the addition for selecting weights forward pushes away It is single to recommend song, it is single to finally obtain recommendation song;

Obtained recommendation song is singly sent to user terminal by the 5th step.

Further, the step two includes following procedure:

Step 201) assigns initial value processing:After server receives brain feedback data focus index x and allowance index y, First determine whether user listens the song for the first time, if listening for the first time, then it is 0.5 to assign its weights, is otherwise taken out at last time Weights after reason are subsequently calculated;

Step 202) normalized:By focus index x and allowance index y normalization, make its be mapped as [0,1] it Between decimal, transfer function is as follows:

Wherein max and min is respectively the maximum and minimum value of focus index x, and x is the value actually measured, and x* is conversion Value afterwards;

It can similarly obtain loosening the normalized value y* of index degree y.

Further, the step three includes following procedure:

Step 301) point is away from calculating:Point (x*, y*) is calculated to the distance between standard point (0.5,0.5) d;

Step 302) right value update:It is said if (x*, y*) is point (0.5,0.5) without updating weights, otherwise divides following feelings Condition:

(1) (x*, y*) is located at first quartile, i.e. x*>0.5 and y*>0.5, which is handled:

W*=w+ (1+d)

Wherein w is the weights before update, and w* is updated weights;

(2) (x*, y*) is located at fourth quadrant, i.e. x*>0.5 and y*<0.5, song weighting processing:

W*=w+ (0.5+d)

Wherein w is the weights before update, and w* is updated weights;

(3) (x*, y*) is located at third quadrant, i.e. x*<0.5 and y*>0.5, which subtracts power processing:

W*=w- (0.5+d)

Wherein w is the weights before update, and w* is updated weights;

(4) (x*, y*) is located at third quadrant, i.e. x*<0.5 and y*<0.5, which subtracts power processing:

W*=w- (1+d)

Wherein w is the weights before update, and w* is updated weights;

Judge whether w* crosses the border after having updated weights, even w*>10, then enable w*=100;If w*<- 100, then enable w*=- 100。

Advantageous effect:What the present invention realized has technical effect that, brain wave changes when listening different songs according to user Trend is analyzed, and the helpful music of concentration when being read to user is selected to promote to use with proposed algorithm auxiliary Person reads the efficiency with study so that music recommendation no longer sticks to the traditional approach such as the recommendation of keyword, makes recommendation more Precisely.The music that can be adapted for the various states such as work, study and sleep is recommended, and a more human nature is created that for user Change and targetedly man-machine system.

Description of the drawings

Fig. 1 is the method flow diagram of present system.

Fig. 2 is the digital independent flow chart of the present invention.

Fig. 3 is the algorithm designing points figure of the present invention.

Specific implementation mode

The present invention is further described below in conjunction with the accompanying drawings.

The present invention includes two main contents:Dynamically personalized song recommendations system based on brain wave earphone, second is that acquisition The current EEG signals of user judge the focus and allowance of user, to the method for current song assignment and sequence, the method Core is the algorithm of song recommendations system.

Such as attached drawing 1, the dynamically personalized song recommendations system based on brain wave earphone, system includes brain wave earphone system, uses Family terminal and server, the brain feedback data that the brain wave earphone system built-in algorithm obtains are transmitted to user terminal, user The brain feedback data received is transmitted to server after server receives brain feedback data and passes through what is improved by terminal Song recommendations algorithm carries out brain feedback data to assign weights processing, is resequenced, is formed to song according to updated weights Song patrilineal line of descent with only one son in each generation is recommended to transport to user terminal.

The method and step that system is realized is as follows:

Step 1:User wears brain wave earphone, reads the eSense focus index x (β obtained by earphone built-in algorithm Wave) and eSense allowance indexes y (α waves).

Step 2:The data read are sent to user terminal by bluetooth.

Step 3:The data read are uploaded to PC server by user terminal by network communication.

Step 4:After server receives data, EEG signals are handled by the song recommendations algorithm improved, are counted Calculation obtains sampled value, and weights are assigned to song, forms a individualized music database.

Step 5:Server is sent to user terminal by network communication, by obtained musical database, i.e., online song is single, Realize personalized choosing song function.

(2) digital independent and the algorithm of song recommendations system, are shown in attached drawing 2 and attached drawing 3.

EEG signals data of the user when listening to music are fitted to several spies by brain wave earphone by built-in eSense algorithms Levy waveform, eSense focus indexes x (β waves) is consequently formed, eSense allowance indexes y (α waves) as shown in table 1 can be with needle Fluctuation tendency of different users' eeg signal in normal range (NR) and individual difference are compensated into Mobile state.

1 brain wave earphone system built-in algorithm data of table

Waveform catalog Frequency The state of mind α waves 8-13Hz Relaxation state but not sleepy, meditation, it is conscious, it is tranquil

Low frequency β waves 12-15Hz The kinaesthesis rhythm and pace of moving things, i.e., it is light and absorbed, there is harmony Intermediate frequency β waves 16Hz-20Hz Thinking, for self and ambient enviroment Clear consciousness High frequency β waves 21Hz-30Hz Vigilance, excitement

The present invention chooses eSense focus index x and is combined with allowance index y, since the index is dynamic data Stream, for extraction Long-term change trend information in time, this song recommendations algorithm is required according to generating date, and has processing high amount of traffic Ability.

The recommendation process based on brain wave earphone carried out using the present invention is as follows:

Step 1, brain wave earphone is worn in subject's head, forehead sensor is affixed on left front volume, by ear loop Ear clip is clipped on ear-lobe, it is ensured that sensor is completely attached to forehead, ear clip and ear-lobe.

Step 2, eeg data is sent to client terminal after built-in algorithm is handled by bluetooth.

Step 3, the blue-teeth data read is uploaded to server by client terminal.

Step 4, server receives data and (after focus index x, allowance index y), such as attached drawing 2, first determines whether user Whether listen the song for the first time, if listening for the first time, then it is 0.5 to assign its weights, the weights that otherwise take out that treated last time into Row is follow-up to be calculated.

Step 5, focus index x and allowance index y are normalized, i.e., linear transformation is done to initial data, it is made to reflect The decimal between [0,1] is penetrated, transfer function is as follows:

Wherein max and min is respectively the maximum and minimum value of focus index x, and x is actual measured value, and x* is after converting Value.The normalized value y* that allowance index y can similarly be obtained is conducive to subsequently after initial data is done normalized It calculates.

Step 6, point (x*, y*) is calculated to the distance between standard point (0.5,0.5) d, which will be used as update weight When one of Main Basiss.

Step 7, illustrate that user is moderate to the song focus and allowance army if (x*, y*) is point (0.5,0.5), Both it did not liked or did not disliked, there is no need to update weights.If not, then to divide and discuss following aspects, such as attached drawing 3:

(1) (x*, y*) is located at first quartile, i.e. x*>0.5 and y*>0.5, illustrate that user has significantly when listening the song Loosen sense i.e. focus, therefore adds some upper weights to the song:

W*=w+ (1+d)

Wherein w is the weights before update, and w* is updated weights.

(2) (x*, y*) is located at fourth quadrant, i.e. x*>0.5 and y*<0.5, illustrate that user has significantly when listening the song Focus (this be user read a book learn when most important factor) but not enough loosen, therefore add some less to the song and weigh Weight:

W*=w+ (0.5+d)

Wherein w is the weights before update, and w* is updated weights.

(3) (x*, y*) is located at third quadrant, i.e. x*<0.5 and y*>0.5, illustrate although user compares when listening the song Loosen but be not enough absorbed in, therefore subtracts some weights less to the song:

W*=w- (0.5+d)

Wherein w is the weights before update, and w* is updated weights.

(4) (x*, y*) is located at third quadrant, i.e. x*<0.5 and y*<0.5, it is clear that this be reading this be most not desired to the feelings seen Condition is neither absorbed in nor is loosened, therefore subtracts some weights to the song more:

W*=w- (1+d)

Wherein w is the weights before update, and w* is updated weights.

Judge whether w* crosses the border after having updated weights, even w*>10, then enable w*=100;If w*<- 100, then enable w*=- 100 to prevent the excessive or too small of data.

Step 8, song is ranked up according to above-mentioned steps calculated new weights, the addition for selecting weights forward pushes away It is single to recommend song, it is single to finally obtain recommendation song.

The above is only a preferred embodiment of the present invention, it should be pointed out that:For the ordinary skill people of the art For member, various improvements and modifications may be made without departing from the principle of the present invention, these improvements and modifications are also answered It is considered as protection scope of the present invention.

Claims (1)

1. a kind of method of the dynamically personalized song recommendations system based on brain wave earphone, which is characterized in that described based on brain electricity The dynamically personalized song recommendations system of earphone includes brain wave earphone system, user terminal and server;The method comprising the steps of:
The first step:The user terminal reads the brain feedback data that brain wave earphone system-computed obtains by bluetooth, respectively Focus index x and allowance index y, and it is uploaded to server;
Second step:The brain feedback data focus index x and allowance index y that the server receives the first step return One change is handled, including step 201) is to 203):
Step 201) assigns initial value processing:After server receives brain feedback data focus index x and allowance index y, first Judge whether user listens the song for the first time, if listening for the first time, then it is 0.5 to assign its weights, after otherwise taking out last time processing Weights subsequently calculated;
Step 202) normalized:By focus index x and the y normalization of allowance index, it is made to be mapped as between [0,1] Decimal, transfer function are as follows:
Wherein max and min is respectively the maximum and minimum value of focus index x, and x is the value actually measured, and x* is transformed Value;
Step 203) can similarly obtain the normalized value y* of allowance index y;
Third walks:By calculate second step normalization after result calculate point away from, according to different situations carry out right value update processing, Including step:
Step 301) point is away from calculating:Point (x*, y*) is calculated to the distance between standard point (0.5,0.5) d;
Step 302) right value update:Illustrate, without updating weights, otherwise to divide following situation if (x*, y*) is point (0.5,0.5):
(1) (x*, y*) is located at first quartile, i.e. x*>0.5 and y*>0.5, which is handled:
W*=w+ (1+d)
Wherein w is the weights before update, and w* is updated weights;
(2) (x*, y*) is located at fourth quadrant, i.e. x*>0.5 and y*<0.5, song weighting processing:
W*=w+ (0.5+d)
Wherein w is the weights before update, and w* is updated weights;
(3) (x*, y*) is located at third quadrant, i.e. x*<0.5 and y*>0.5, which subtracts power processing:
W*=w- (0.5+d)
Wherein w is the weights before update, and w* is updated weights;
(4) (x*, y*) is located at third quadrant, i.e. x*<0.5 and y*<0.5, which subtracts power processing:
W*=w- (1+d)
Wherein w is the weights before update, and w* is updated weights;
Judge whether w* crosses the border after having updated weights, even w*>100, then enable w*=100;If w*<- 100, then enable w*=-100;
4th step:Calculated new weights are walked according to third to be ranked up song, song is recommended in the addition for selecting weights forward It is single, it is single to finally obtain recommendation song;
Obtained recommendation song is singly sent to user terminal by the 5th step.
CN201510467034.6A 2015-08-03 2015-08-03 A kind of dynamically personalized song recommendations system and method based on brain wave earphone CN105045898B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510467034.6A CN105045898B (en) 2015-08-03 2015-08-03 A kind of dynamically personalized song recommendations system and method based on brain wave earphone

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510467034.6A CN105045898B (en) 2015-08-03 2015-08-03 A kind of dynamically personalized song recommendations system and method based on brain wave earphone

Publications (2)

Publication Number Publication Date
CN105045898A CN105045898A (en) 2015-11-11
CN105045898B true CN105045898B (en) 2018-07-20

Family

ID=54452445

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510467034.6A CN105045898B (en) 2015-08-03 2015-08-03 A kind of dynamically personalized song recommendations system and method based on brain wave earphone

Country Status (1)

Country Link
CN (1) CN105045898B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107423352A (en) * 2017-05-25 2017-12-01 杭州回车电子科技有限公司 Music recommends method and system
CN107402635A (en) * 2017-07-31 2017-11-28 天津易念波科技有限公司 With reference to brain wave and the mental health adjusting method and system of virtual reality
CN109714669A (en) * 2019-01-15 2019-05-03 浙江强脑科技有限公司 Method for playing music, earphone and computer readable storage medium

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101467875A (en) * 2007-12-28 2009-07-01 周常安 Ear-wearing type physiology feedback device
CN102446533A (en) * 2010-10-15 2012-05-09 盛乐信息技术(上海)有限公司 Music player
CN103412646A (en) * 2013-08-07 2013-11-27 南京师范大学 Emotional music recommendation method based on brain-computer interaction

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101467875A (en) * 2007-12-28 2009-07-01 周常安 Ear-wearing type physiology feedback device
CN102446533A (en) * 2010-10-15 2012-05-09 盛乐信息技术(上海)有限公司 Music player
CN103412646A (en) * 2013-08-07 2013-11-27 南京师范大学 Emotional music recommendation method based on brain-computer interaction

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
改变音乐节拍对脑电功率谱变异的影响极其心理意义;张德乾等;《绵阳师范学院学报》;20140228;第33卷(第2期);全文 *

Also Published As

Publication number Publication date
CN105045898A (en) 2015-11-11

Similar Documents

Publication Publication Date Title
Sterne The mp3 as cultural artifact
TWI487385B (en) Volume adjusting device and adjusting method of the same
US20060079291A1 (en) Method and apparatus for multi-sensory speech enhancement on a mobile device
US10321842B2 (en) System and method for associating music with brain-state data
JP2010092065A (en) Localized audio networks and associated digital accessories
US20130243227A1 (en) Personal communication device with hearing support and method for providing the same
James et al. Adaptive dynamic range optimization for cochlear implants: a preliminary study
US20010005420A1 (en) Optimum solution method, hearing aid fitting apparatus utilizing the optimum solution method, and system optimization adjusting method and apparatus
EP2229228B1 (en) System and method for automatically creating an atmosphere suited to social setting and mood in an environment
Luo et al. Cochlear implants special issue article: Vocal emotion recognition by normal-hearing listeners and cochlear implant users
JP6471101B2 (en) Information processing system and information processing apparatus
Yu et al. Modeling subjective evaluation of soundscape quality in urban open spaces: An artificial neural network approach
US9609442B2 (en) Smart hearing aid
TWI584272B (en) Method and system for self-managed sound enhancement
CN109076284A (en) The voice control of media playback system
US20100208631A1 (en) Inaudible methods, apparatus and systems for jointly transmitting and processing, analog-digital information
CN105592777A (en) Method and system for sleep management
US20160234606A1 (en) Method for augmenting hearing
JP2007213401A (en) Community site server and program for configurating community based on user preferable music data
CN104361016B (en) A kind of method and device that music effect is adjusted according to motion state
CN101904151A (en) Method of controlling communications between at least two users of a communication system
US6875167B2 (en) Apparatus for generating brain wave-inducing signals and network system including the same
CN103680545B (en) Audio playback system and method for controlling playback
US9652532B2 (en) Methods for operating audio speaker systems
JP2018512607A (en) Method, system and medium for correction of environmental background noise based on mood and / or behavior information

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
CP02 Change in the address of a patent holder
CP02 Change in the address of a patent holder

Address after: 210093 Nanjing University Science Park, 22 Hankou Road, Gulou District, Nanjing City, Jiangsu Province

Patentee after: Southeast University

Address before: 211103 No. 59 Wan'an West Road, Dongshan Street, Jiangning District, Nanjing City, Jiangsu Province

Patentee before: Southeast University