WO2023035647A1 - 基于可穿戴设备的音乐推荐方法、装置、设备及存储介质 - Google Patents

基于可穿戴设备的音乐推荐方法、装置、设备及存储介质 Download PDF

Info

Publication number
WO2023035647A1
WO2023035647A1 PCT/CN2022/092080 CN2022092080W WO2023035647A1 WO 2023035647 A1 WO2023035647 A1 WO 2023035647A1 CN 2022092080 W CN2022092080 W CN 2022092080W WO 2023035647 A1 WO2023035647 A1 WO 2023035647A1
Authority
WO
WIPO (PCT)
Prior art keywords
music
recommended
relaxation
target
target user
Prior art date
Application number
PCT/CN2022/092080
Other languages
English (en)
French (fr)
Inventor
汪孔桥
俞轶
朱国康
张聪
孟孜
Original Assignee
安徽华米健康科技有限公司
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 安徽华米健康科技有限公司 filed Critical 安徽华米健康科技有限公司
Publication of WO2023035647A1 publication Critical patent/WO2023035647A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/60—Information retrieval; Database structures therefor; File system structures therefor of audio data
    • G06F16/63—Querying
    • G06F16/635—Filtering based on additional data, e.g. user or group profiles
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning

Definitions

  • the present disclosure relates to the field of computer technology, and in particular to a music recommendation method, device, device and storage medium based on wearable devices.
  • the traditional sleep aid music recommendation method relies on the user's interactive behavior or subjective preference selection, which has subjective uncertainty. For example, the songs on a certain playlist are only played in a loop, and the order in which they are played is not absolutely related to the effect of sleep aids. Therefore, how to achieve accurate personalized sleep aid music recommendations is a major technical challenge.
  • the present disclosure provides a wearable device-based music recommendation method, device, device and storage medium.
  • a method for recommending music based on a wearable device including:
  • At least one target recommended music is determined according to the target user's current relaxation state and multiple relaxation parameters corresponding to the music to be recommended, including:
  • the second number is greater than the first number.
  • the relaxation state evaluation model generated through training before inputting the physiological parameters of the target user into the relaxation state evaluation model generated through training, it also includes:
  • the first training data set includes physiological parameters of multiple reference users and relaxation states marked according to the EEG data of the multiple reference users;
  • the initial relaxation state evaluation model is corrected to generate the relaxation state evaluation model generated through training.
  • the method before determining at least one target recommended music according to the current relaxation state of the target user and the relaxation parameters corresponding to multiple music to be recommended, the method further includes:
  • attribute information, historical sleep data and current relaxation state of the target user determine the target music style currently corresponding to the target user
  • candidate music belonging to the target music style is acquired from a candidate music library as the multiple pieces of music to be recommended.
  • the method before determining at least one target recommended music according to the current relaxation state of the target user and the relaxation parameters corresponding to multiple music to be recommended, the method further includes:
  • the method also includes:
  • the second training data set includes a first relaxation curve when playing the plurality of music to be recommended according to a plurality of reference users and a second relaxation curve when the music to be recommended is not played ;
  • the initial relaxation parameter estimation model is corrected to generate the relaxation parameter estimation model generated after training.
  • the playing the at least one target recommended music includes:
  • the plurality of target recommended music and the fused music are played sequentially.
  • the pairwise fusion of the multiple target recommended music to obtain multiple fused music includes:
  • the playback order from the back to the front, intercept the first music segment with a preset duration from the previous music in every two target recommended music;
  • a second music segment with a preset duration is intercepted from the next music in every two target recommended music
  • the first music segment and the second music segment are fused to obtain fused music.
  • the merging of the first music segment and the second music segment to obtain the fused music includes:
  • the first weight sequence and the second weight sequence respectively include a plurality of weight values, and the plurality of first weight values in the first weight sequence gradually decrease, and the plurality of first weight values in the second weight sequence The two weight values increase gradually, and the sum of each first weight value and the corresponding second weight value is 1.
  • a wearable device-based music recommendation device including:
  • An acquisition module configured to acquire the physiological parameters of the target user collected by the wearable device
  • the first determination module is used to input the physiological parameters of the target user into the relaxation state evaluation model generated through training, so as to determine the current relaxation state of the target user;
  • the second determining module is configured to determine at least one target recommended music according to the current relaxation state of the target user and relaxation parameters corresponding to multiple music to be recommended, wherein the target recommended music is used to play to the target user.
  • an electronic device including:
  • the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor, so that the at least one processor can execute the method described in the above one aspect embodiment.
  • a non-transitory computer-readable storage medium storing computer instructions, on which a computer program is stored, and the computer instructions are used to enable the computer to execute the above-mentioned embodiment of the first aspect.
  • a computer program product including a computer program, when the computer program is executed by a processor, the method described in the above-mentioned one embodiment is implemented
  • the device first obtains the physiological parameters of the target user collected by the wearable device, and inputs the physiological parameters of the target user into the relaxation state evaluation model generated through training to determine the current relaxation state of the target user, and then according to The target user's current relaxation state and the relaxation parameters corresponding to the plurality of music to be recommended determine at least one target recommended music, and the target recommended music is used to play to the target user. Therefore, by wearing a wearable device on the target user, the relaxation state can be evaluated in real time, and this method takes into account the portability of the wearable device and the fitting ability of machine learning technology, and can accurately and effectively provide target users with Music that brings a relaxing effect.
  • the device first acquires the first training data set, wherein the first training data set includes the physiological parameters of multiple reference users and the relaxation state marked according to the EEG data of multiple reference users, and the reference
  • the user's physiological parameters are input into the initial relaxation state evaluation model to obtain the predicted relaxation state, and then according to the difference between the marked relaxation state and the predicted relaxation state, the initial relaxation state evaluation model is corrected to generate the training-generated Relaxation state evaluation model, and then obtain the physiological parameters of the target user collected by the wearable device, and then input the physiological parameters of the target user into the relaxation state evaluation model generated after training to determine the current relaxation state of the target user, and then according to the target
  • the user's current relaxation state and the relaxation parameters corresponding to the plurality of music to be recommended determine at least one target recommended music, and the target recommended music is used to play to the target user. Therefore, by using the reference user to determine the training data set and modifying the relaxation state evaluation model, the relaxation parameters of each music to be recommended can be determined more accurately, so that the target user can be
  • the device first acquires the physiological parameters of the target user collected by the wearable device, and inputs the physiological parameters of the target user into the relaxation state evaluation model generated through training to determine the current relaxation state of the target user, Then, according to the current relaxation state of the target user and the relaxation parameters corresponding to a plurality of music to be recommended, at least one target recommended music is determined, and then multiple target recommended music is fused in pairs to obtain multiple fused music, and according to each Two target recommended music corresponding to one fused music, determine the playing sequence of multiple target recommended music and fused music, and finally play multiple target recommended music and fused music sequentially based on the playing sequence.
  • FIG. 1 is a schematic flow diagram of a music recommendation method based on a wearable device according to the present disclosure
  • FIG. 2 is a schematic flowchart of another music recommendation method based on a wearable device according to the present disclosure
  • FIG. 3 is a schematic flowchart of another music recommendation method based on a wearable device according to the present disclosure
  • FIG. 4 is a structural block diagram of a music recommendation device based on a wearable device provided by the present disclosure
  • Fig. 5 is a structural block diagram of an electronic device provided by the present disclosure.
  • the music recommendation method based on a wearable device proposed in this disclosure can be executed by the music recommendation device based on a wearable device provided in this disclosure, and can also be executed by an electronic device provided in this disclosure, wherein the electronic device can include but not limited to a desktop computer, a tablet A terminal device such as a computer can also be a server.
  • the device for generating sleep-aid music provided by the present disclosure will be used to execute a music recommendation method based on a wearable device provided by the present disclosure. "Device”.
  • Fig. 1 is a schematic flowchart of a music recommendation method based on a wearable device according to an embodiment of the present disclosure. As shown in FIG. 1 , the music recommendation method based on a wearable device may include the following steps S101 to S104.
  • Step S101 acquiring the physiological parameters of the target user collected by the wearable device.
  • the wearable device can be worn by the target user, and the target user's physiological parameters can be obtained by contacting or pressing against the target user's skin through the sensor.
  • the physiological parameters of the target user may be the target user's heart rate, respiration rate, heart rate variability, respiration rate variability, skin electrical impedance, blood pressure, blood oxygen, and sleep stages, etc.
  • the physiological parameters of the target user can be used as indicators for later evaluating the relaxation state of the target user.
  • the heart rate of the human body will decrease in a relaxed state. According to this prior knowledge, it can be considered that if the heart rate of the human body decreases significantly, it indicates that the human body is in a relaxed state at this time.
  • Step S102 input the physiological parameters of the target user into the relaxation state assessment model generated through training, so as to determine the current relaxation state of the target user.
  • the music recommendation for stress relief and sleep aid relies on the accurate assessment of the target user's own physical relaxation state. Therefore, after the device acquires the physiological parameters of the target user in real time, it needs to input the physiological parameter data into the pre-trained model to estimate the current relaxation state of the target user.
  • the device can generate a label corresponding to the target user's relaxation state through the relaxation state evaluation model, such as "relaxed”, “relatively relaxed”, “not relaxed”, and “stressed”.
  • Step S103 Determine at least one target recommended music according to the current relaxation state of the target user and the relaxation parameters corresponding to the plurality of music to be recommended.
  • the target recommended music may be music that can assist the target user to relax and sleep while considering the target user's own characteristics.
  • the device determines at least one target recommended music according to the current relaxation state of the target user and the relaxation parameters corresponding to multiple music to be recommended, there may be multiple ways.
  • a plurality of music to be recommended corresponds to a relaxation parameter, wherein the size of the relaxation parameter is used to represent the relaxation effect of the music to be recommended.
  • the relaxation parameter is directly proportional to the relaxation performance of the music to be recommended, it means that the larger the relaxation parameter, the higher the relaxation performance of the music to be recommended, that is, the more beneficial it is for the target user to relax, decompress, and help sleep.
  • the same music may have different relaxation parameters for different users.
  • the device can select certain Quantity of music recommended to target users.
  • the top M/4 music tracks can be recommended as the target recommended music to the target user according to the relaxation parameter in descending order.
  • the music track of the first 3M/4 can be selected therefrom as the target recommended music to the target user.
  • Step S104 playing the target recommended music to the target user.
  • the device can play the target recommended music through the playback device. It can be understood that while playing the target recommended music, the device can also use the wearable device to collect the physiological parameters of the target user in real time, so that the relaxation parameters of the music can be dynamically adjusted afterwards.
  • the device first acquires the physiological parameters of the target user collected by the wearable device, and inputs the physiological parameters of the target user into the relaxation state evaluation model generated through training to determine the current relaxation state of the target user, and then according to the target user
  • the user's current relaxation state and the relaxation parameters corresponding to multiple music to be recommended determine at least one target recommended music, and the target recommended music is used to play to the target user. Therefore, by wearing a wearable device on the target user, the relaxation state can be evaluated in real time, and this method takes into account the portability of the wearable device and the fitting ability of machine learning technology, and can accurately and effectively provide target users with Music that brings a relaxing effect.
  • Fig. 2 is a schematic flowchart of a music recommendation method based on a wearable device according to another embodiment of the present disclosure.
  • the music recommendation method based on wearable devices may include the following steps S201 to S207.
  • Step S201 acquiring a first training data set, wherein the first training data set includes physiological parameters of multiple reference users and relaxation states marked according to EEG data of multiple reference users.
  • the first training data set may be collected from reference users.
  • the device can collect the physiological parameter data of the reference user using the wearable device in real time, such as heart rate, respiratory rate, skin electrical impedance, blood pressure, blood oxygen and other physiological parameters.
  • the device It is also possible to collect the EEG data corresponding to each reference user through the EEG acquisition device.
  • the EEG data can be obtained by detecting the EEG signals on the human scalp with an EEG acquisition device, and collecting and processing the EEG signals through related equipment.
  • EEG is a standard that can directly and accurately measure the relaxation state of the human body
  • the present disclosure can determine the marked relaxation state of each reference user according to the EEG data of each reference user.
  • marked relaxation states there may be various kinds of marked relaxation states, such as relaxed, relatively relaxed, slightly relaxed, not relaxed, and the like.
  • the EEG data may include reference user's EEG signals, such as alpha waves, beta waves, theta waves, peak waves, sleep spindles, and the like.
  • the device can determine the marked relaxation state of the reference user. For example, if the parameters of the ⁇ wave feature of a certain reference user are 8 to 13 Hz and the amplitude is 20 to 100 ⁇ V, it can be determined that the state of the current user is "relaxed".
  • Step S202 inputting the physiological parameters of the reference user into the initial relaxation state assessment model to obtain a predicted relaxation state.
  • the initial relaxation state evaluation model may be a network model that has not yet reached a usable state, and the device can then modify and train its network parameters to make it reach a usable state.
  • the reference user's relaxation state is predicted by inputting the reference user's physiological parameters into the initial relaxation state evaluation model, wherein the marked relaxation state can be relaxed, relatively relaxed, slightly relaxed, not relaxed, etc. kind.
  • Step S203 according to the difference between the labeled relaxation state and the predicted relaxation state, the initial relaxation state evaluation model is corrected to generate a trained relaxation state evaluation model.
  • the device can compare the labeled relaxed state with the predicted relaxed state to determine the difference between the labeled relaxed state and the predicted relaxed state.
  • gradient descent stochastic gradient descent, etc. can be used to determine the corrected gradient, and then In this way, the initial evaluation model is corrected to generate a relaxed state evaluation model generated through training.
  • the manner of modifying the gradient in the present disclosure can be determined according to actual needs.
  • Step S204 acquiring the physiological parameters of the target user collected by the wearable device.
  • Step S205 inputting the physiological parameters of the target user into the relaxation state evaluation model generated through training, so as to determine the current relaxation state of the target user.
  • steps S204 and S205 may refer to the above-mentioned embodiments, and details are not described here.
  • Step S206 Determine at least one target recommended music according to the current relaxation state of the target user and relaxation parameters corresponding to multiple music to be recommended, wherein the target recommended music is used to play to the target user.
  • the device determines at least one target recommended music according to the current relaxation state of the target user and the relaxation parameters corresponding to the multiple music to be recommended, it also needs to acquire multiple relaxation parameters corresponding to the music to be recommended.
  • the music features of each music to be recommended and the attribute information of each reference user may be input into a pre-trained relaxation parameter estimation model to determine the relaxation parameters of each music to be recommended.
  • a pre-trained relaxation parameter estimation model to determine the relaxation parameters of each music to be recommended.
  • the second training data set includes a first relaxation curve when playing the plurality of music to be recommended according to a plurality of reference users and a second relaxation curve when the music to be recommended is not played;
  • the initial relaxation parameter estimation model is corrected to generate a relaxation parameter estimation model generated after training.
  • both the first relaxation curve and the second relaxation curve can be represented as function images of time, for example, taking time as the abscissa and fatigue degree as the ordinate, or taking time as the abscissa and relaxation degree as the ordinate. It should be noted that the relaxation curves of the reference user when playing music of different styles may be the same or different.
  • the music feature may be music rhythm, music interval, harmony, pitch and so on. It should be noted that, before inputting the musical features into the initial relaxation parameter prediction model, the musical features of each piece of music may be quantified.
  • the attribute information of the reference user may be information such as age, gender, weight, and music preference of the reference user.
  • the relaxation parameter is also the relaxation effect corresponding to the music, and the relaxation parameters of different styles of music are different.
  • the same music has different relaxation parameters for different target users.
  • the device can mark each music to be recommended by calculating the integral difference in time between the first relaxation curve and the second relaxation curve Relaxation parameters for each reference user.
  • the initial relaxation parameter estimation model may be a network model that has not yet reached a usable state, and the device can then correct and train its network parameters to make it reach a usable state.
  • the predicted relaxation parameters of each music to be recommended for each reference user are obtained .
  • the device can compare the labeled relaxation parameters with the predicted relaxation parameters to determine the difference between the labeled relaxation parameters and the predicted relaxation parameters, for example, gradient descent, stochastic gradient descent, etc. can be used to determine The gradient is corrected, and then the initial relaxed parameter prediction model is corrected to generate a relaxed parameter prediction model generated after training.
  • the method of correcting the gradient in the present disclosure can be determined according to actual needs.
  • the device can determine the relaxation parameters corresponding to multiple music to be recommended by inputting music characteristics of each music to be recommended and attribute information of each target user.
  • the relaxation state of the target user in the previous time window may also be input. For example, if 4 hours is used as a time window, the relaxation state of the target user 4 hours before the current time may be input.
  • the device may determine at least one target recommended music according to the current relaxation state of the target user and the relaxation parameters corresponding to the plurality of music to be recommended.
  • Step S207 playing the target recommended music to the target user.
  • step S207 may refer to the foregoing embodiments, and details are not described here.
  • the device first obtains the first training data set, wherein the first training data set includes the physiological parameters of multiple reference users and the relaxation state marked according to the EEG data of multiple reference users, and the physiological parameters of the reference users Input the initial relaxation state evaluation model to obtain the predicted relaxation state, and then modify the initial relaxation state evaluation model according to the difference between the marked relaxation state and the predicted relaxation state to generate a trained relaxation state evaluation model , and then obtain the physiological parameters of the target user collected by the wearable device, and then input the physiological parameters of the target user into the relaxation state evaluation model generated through training to determine the current relaxation state of the target user, and then according to the current relaxation state of the target user
  • the state and relaxation parameters corresponding to multiple music to be recommended determine at least one target recommended music, and the target recommended music is used to play to the target user. Therefore, by using the reference user to determine the training data set and modifying the relaxation state evaluation model, the relaxation parameters of each music to be recommended can be determined more accurately, so that the target user can be provided with relaxation effects more accurately and effectively. music
  • Fig. 3 is a schematic flowchart of a music recommendation method based on a wearable device according to yet another embodiment of the present disclosure.
  • the wearable device-based music recommendation method may include the following steps S301 to S306.
  • Step S301 acquiring the physiological parameters of the target user collected by the wearable device.
  • Step S302 inputting the physiological parameters of the target user into the relaxation state assessment model generated through training, so as to determine the current relaxation state of the target user.
  • steps S301 and S302 can refer to any of the above-mentioned embodiments, and the present disclosure will not repeat them here.
  • Step S303 Determine at least one target recommended music according to the current relaxation state of the target user and the relaxation parameters corresponding to multiple music to be recommended.
  • the music to be recommended can be candidate music preset in the candidate music library, and can have many different styles.
  • the device Before determining at least one target recommended music, the device can acquire part of the music to be recommended. The following disclosure will Some implementation methods for acquiring music to be recommended are described.
  • the device may first obtain the target user's music preference, attribute information, and historical sleep data, and then determine the target user's current sleep status according to the target user's music preference, attribute information, historical sleep data, and current relaxation state.
  • the target music style, and then according to the music styles to which the plurality of music to be recommended belong, the candidate music belonging to the target music style is obtained from the candidate music library as the plurality of music to be recommended.
  • the sleep aid music suitable for different users is also ever-changing.
  • the device can determine the target music style currently corresponding to the target user, and obtain candidate music belonging to the target music style required by the current user from the candidate music library as a plurality of music to be recommended.
  • the music to be recommended may be multiple pieces of music.
  • step S304 multiple target recommended musics are fused in pairs to obtain multiple fused musics.
  • weighted fusion may be performed on every two adjacent target recommended music according to the playing sequence of the multiple target recommended music, so as to obtain multiple fused music.
  • the device can also firstly intercept the first music segment with a preset duration from the previous music in every two target recommended music according to the playing order from back to front, and then according to the playing order, from front to front Finally, a second music segment with a preset duration is intercepted from the last music in every two target recommended music.
  • the device can determine the playing order of the target music recommendations in descending order of relaxation parameters, and then play each two Neighboring target recommendation music for fusion.
  • the device may combine the target recommended music in pairs into (A, B) and (B, C).
  • the device can select the first music segment A1 with a preset duration (for example, 20s) from music A from back to front, and select from music B from front to back The second music segment B1 of 20s.
  • first music segment and the second music segment are fused to obtain a plurality of fused music. It should be understood that there may be multiple ways to fuse the first music segment and the second music segment.
  • each sub-segment of the first music segment and the second music segment may be added separately, so that the fused music contains elements of the two target recommended music.
  • the device can also fuse each sub-segment of the first music segment and the second music segment based on the first weight sequence and the second weight sequence to generate fused music.
  • the first weight sequence and the second weight sequence respectively contain a plurality of weight values, and the plurality of first weight values in the first weight sequence gradually decrease, and the plurality of second weight values in the second weight sequence gradually increase, and every The sum of each first weight value and the corresponding second weight value is 1.
  • the first weight sequence is a weight sequence corresponding to each sub-segment of the first music segment
  • the second weight sequence is a weight sequence corresponding to each sub-segment of the second music segment.
  • the first music segment is A1
  • A1 includes sub-segments a1, a2, a3
  • the second music segment is B1, wherein B1 includes sub-segments b1, b2, b3.
  • the weights corresponding to the first weight sequence a1, a2, a3 are 0.8, 0.5, 0.3 respectively
  • the weights corresponding to the second weight sequence b1, b2, b3 are 0.2, 0.5, 0.7 respectively.
  • the proportion of the first music segment A1 is reduced, and the proportion of the second music segment B1 is increased.
  • Step S305 according to the two target recommended music corresponding to each fused music, determine the playing sequence of multiple target recommended music and fused music.
  • the present disclosure in order to avoid music switching caused by different target recommended music during sequential playback, causing sensory changes and causing interference to users. After two-by-two fusion of multiple target recommended music, elements of two target recommended music can be contained in the generated fusion music. Therefore, during playback, the fused music can be set to be played between the corresponding two target recommended music, so that the music switching process can be made insensitive, and the user experience is improved.
  • the device can set the music generated by the fusion of A and B to be played between A and B, and set the music of B and C to be played between Play between B and C, and so on.
  • the similarity between the target recommended music can be improved, the user can avoid the abrupt feeling of music switching, and more relaxing songs can be generated in real time to enrich the music library.
  • Step S306 based on the play sequence, play multiple target recommended music and the fused music in sequence.
  • the device can use the playing device to play the multiple target recommended music and the fused music in a playing order.
  • the wearable device can also collect the physiological parameters of the target user in real time, so that the recommended music can be adjusted according to the real-time relaxation state of the target user.
  • the device first acquires the physiological parameters of the target user collected by the wearable device, and inputs the physiological parameters of the target user into the relaxation state evaluation model generated through training to determine the current relaxation state of the target user, and then according to the target user Based on the user's current relaxation state and the relaxation parameters corresponding to multiple music to be recommended, at least one target music recommendation is determined, and then multiple target music recommendations are fused in pairs to obtain multiple fused music, and according to each fused music The two target recommended music corresponding to the music of the music, determine the playing sequence of the multiple target recommended music and the fused music, and finally play the multiple target recommended music and the fused music sequentially based on the playing sequence.
  • Fig. 4 is a structural block diagram of an apparatus for recommending music based on a wearable device according to an embodiment of the present disclosure.
  • the wearable device-based music recommendation device includes: an acquisition module 410 , a first determination module 420 , a second determination module 430 and a playback module 440 .
  • An acquisition module 410 configured to acquire the physiological parameters of the target user collected by the wearable device
  • the first determining module 420 is configured to input the physiological parameters of the target user into the relaxation state assessment model generated through training, so as to determine the current relaxation state of the target user;
  • the second determination module 430 is configured to determine at least one target recommended music according to the current relaxation state of the target user and the relaxation parameters corresponding to a plurality of music to be recommended;
  • the playing module 440 is configured to play the target recommended music to the target user.
  • the first determining module 420 is further configured to:
  • the first training data set includes physiological parameters of multiple reference users and relaxation states marked according to EEG data of multiple reference users;
  • the initial relaxation state evaluation model is corrected to generate the relaxation state evaluation model generated through training.
  • the second determination module 430 is further configured to:
  • attribute information, historical sleep data and current relaxation state of the target user determine the target music style currently corresponding to the target user
  • candidate music belonging to the target music style is acquired from a candidate music library as the multiple pieces of music to be recommended.
  • the second determination module 430 is further configured to:
  • the device further includes:
  • the second obtaining module is used to obtain the second training data set, wherein the second training data set includes the first relaxation curve when playing a plurality of music to be recommended according to a plurality of reference users and when not playing the music to be recommended
  • the third determining module is used to mark the relaxation parameters of each of the music to be recommended for each of the reference users;
  • the third acquisition module is used to input the music features of each of the music to be recommended and the attribute information of each of the reference users into the initial relaxation parameter estimation model, so as to obtain the pair of each of the music to be recommended for each said reference user's predicted relaxation parameters;
  • the correction module is configured to correct the initial relaxation parameter prediction model according to the difference between the marked relaxation parameter and the predicted relaxation parameter, so as to generate the relaxation parameter prediction model generated after training.
  • the playback module includes:
  • a fusion unit configured to fuse the plurality of target recommended music in pairs to obtain a plurality of fused music
  • a determining unit configured to determine the playback order of the plurality of target recommended music and the fused music according to the two target recommended music corresponding to each of the fused music;
  • the playing unit is configured to play the plurality of target recommended music and the fused music sequentially based on the playing order.
  • the fusion unit includes:
  • the first intercepting subunit is used to intercept the first music segment with a preset duration from the previous music in each of the two target recommended music from back to front according to the playback order;
  • the second intercepting subunit is used to intercept the second music segment with a preset duration from the next music in every two target recommended music from front to back according to the playback order;
  • the fusion subunit is configured to fuse the first music segment and the second music segment to obtain fused music.
  • the fusion subunit is specifically used for:
  • the first weight sequence and the second weight sequence respectively include a plurality of weight values, and the plurality of first weight values in the first weight sequence gradually decrease, and the plurality of first weight values in the second weight sequence The two weight values increase gradually, and the sum of each first weight value and the corresponding second weight value is 1.
  • the device first obtains the physiological parameters of the target user collected by the wearable device, and inputs the physiological parameters of the target user into the relaxation state evaluation model generated through training to determine the current relaxation state of the target user, and then according to the target user
  • the current relaxation state and the relaxation parameters corresponding to the plurality of music to be recommended determine at least one target recommended music, and the target recommended music is used to play to the target user. Therefore, by wearing a wearable device on the target user, the relaxation state can be evaluated in real time, and this method takes into account the portability of the wearable device and the fitting ability of machine learning technology, and can accurately and effectively provide target users with Music that brings a relaxing effect.
  • the present disclosure also provides an electronic device, a readable storage medium, and a computer program product.
  • FIG. 5 shows a schematic block diagram of an example electronic device 500 that may be used to implement embodiments of the present disclosure.
  • Electronic device is intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers.
  • Electronic devices may also represent various forms of mobile devices, such as personal digital processing, cellular telephones, smart phones, wearable devices, and other similar computing devices.
  • the above-mentioned electronic device may be a wearable device, and the wearable device is used to collect the physiological parameters of the target user, and the wearable device may be used to determine the target recommended music according to the music recommendation method of the present disclosure, and recommend the target Music is played directly to the target user on the wearable device.
  • the above-mentioned electronic device may be a mobile terminal, the wearable device collects the physiological parameters of the target user, and the wearable device transmits the collected physiological parameters to the mobile terminal (the user of the mobile terminal is usually the same as the user of the wearable device) same person, in some cases the user of the mobile terminal is not the same person as the user of the wearable device), and is handled by an application running on the mobile terminal.
  • the mobile terminal is wirelessly connected to the wearable device.
  • the application running on the mobile terminal can be used to determine the target recommended music according to the music recommendation method of the present disclosure, and play the target recommended music to the target user.
  • the electronic device can be a server
  • the wearable device is used to collect the physiological parameters of the target user
  • the wearable device transmits the collected physiological parameters to the mobile terminal (the user of the mobile terminal is usually the same as the user of the wearable device)
  • the user of the mobile terminal is not the same as the user of the wearable device)
  • there is an application program running on the mobile terminal and the mobile terminal transmits the received physiological parameters to a remote server (for example, a cloud server) , and is processed by a program running on the remote server.
  • the mobile terminal is connected to the wearable device through wireless communication
  • the remote server is connected to the mobile terminal through wireless communication.
  • the remote server can be used to determine the target recommended music according to the music recommendation method of the present disclosure, and push the target recommended music to the mobile terminal, so that the target recommended music can be played on the mobile terminal to the target user.
  • the device 500 includes a computing unit 501 that can execute according to a computer program stored in a read-only memory (ROM) 502 or loaded from a storage unit 508 into a random-access memory (RAM) 503. Various appropriate actions and treatments. In the RAM 503, various programs and data necessary for the operation of the device 500 can also be stored.
  • the computing unit 501, ROM 502, and RAM 503 are connected to each other through a bus 504.
  • An input/output (I/O) interface 505 is also connected to the bus 504 .
  • the I/O interface 505 includes: an input unit 506, such as a keyboard, a mouse, etc.; an output unit 507, such as various types of displays, speakers, etc.; a storage unit 508, such as a magnetic disk, an optical disk, etc. ; and a communication unit 509, such as a network card, a modem, a wireless communication transceiver, and the like.
  • the communication unit 509 allows the device 500 to exchange information/data with other devices over a computer network such as the Internet and/or various telecommunication networks.
  • the computing unit 501 may be various general-purpose and/or special-purpose processing components having processing and computing capabilities. Some examples of computing units 501 include, but are not limited to, central processing units (CPUs), graphics processing units (GPUs), various dedicated artificial intelligence (AI) computing chips, various computing units that run machine learning model algorithms, digital signal processing processor (DSP), and any suitable processor, controller, microcontroller, etc.
  • the calculation unit 501 executes various methods and processes described above, such as a music recommendation method based on a wearable device.
  • a wearable device-based music recommendation method may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as storage unit 508 .
  • part or all of the computer program may be loaded and/or installed on the device 500 via the ROM 502 and/or the communication unit 509 .
  • the computer program When the computer program is loaded into the RAM 503 and executed by the computing unit 501, one or more steps of the above-described wearable device-based music recommendation method can be performed.
  • the computing unit 501 may be configured in any other appropriate way (for example, by means of firmware) to execute a wearable device-based music recommendation method.
  • Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips Implemented in a system of systems (SOC), load programmable logic device (CPLD), computer hardware, firmware, software, and/or combinations thereof.
  • FPGAs field programmable gate arrays
  • ASICs application specific integrated circuits
  • ASSPs application specific standard products
  • SOC system of systems
  • CPLD load programmable logic device
  • computer hardware firmware, software, and/or combinations thereof.
  • programmable processor can be special-purpose or general-purpose programmable processor, can receive data and instruction from storage system, at least one input device, and at least one output device, and transmit data and instruction to this storage system, this at least one input device, and this at least one output device an output device.
  • Program codes for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special purpose computer, or other programmable data processing devices, so that the program codes, when executed by the processor or controller, make the functions/functions specified in the flow diagrams and/or block diagrams Action is implemented.
  • the program code may execute entirely on the machine, partly on the machine, as a stand-alone software package partly on the machine and partly on a remote machine or entirely on the remote machine or server.
  • a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device.
  • a machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium.
  • a machine-readable medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing.
  • machine-readable storage media would include one or more wire-based electrical connections, portable computer discs, hard drives, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or flash memory), optical fiber, compact disk read only memory (CD-ROM), optical storage, magnetic storage, or any suitable combination of the foregoing.
  • RAM random access memory
  • ROM read only memory
  • EPROM or flash memory erasable programmable read only memory
  • CD-ROM compact disk read only memory
  • magnetic storage or any suitable combination of the foregoing.
  • the systems and techniques described herein can be implemented on a computer having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) display device) for displaying information to the target user. monitor); and a keyboard and pointing device (eg, a mouse or trackball) through which the intended user can provide input to the computer.
  • a display device e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) display device
  • keyboard and pointing device eg, a mouse or trackball
  • Other types of devices may also be used to provide interaction with the target user; for example, the feedback provided to the target user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and may be in any form (including acoustic input, speech input or, tactile input) to receive input from the target user.
  • the systems and techniques described herein can be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., as a a user computer having a graphical user interface or a web browser through which an intended user may interact with an embodiment of the systems and techniques described herein), or including such backend components, middleware components , or any combination of front-end components in a computing system.
  • the components of the system can be interconnected by any form or medium of digital data communication, eg, a communication network. Examples of communication networks include: local area networks (LANs), wide area networks (WANs), the Internet, and blockchain networks.
  • a computer system may include clients and servers.
  • Clients and servers are generally remote from each other and typically interact through a communication network.
  • the relationship of client and server arises by computer programs running on the respective computers and having a client-server relationship to each other.
  • the server can be a cloud server, also known as cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the problem of traditional physical host and VPS service ("Virtual Private Server", or "VPS”) Among them, there are defects such as difficult management and weak business scalability.
  • the server can also be a server of a distributed system, or a server combined with a blockchain.
  • the device first acquires the physiological parameters of the target user collected by the wearable device, and inputs the physiological parameters of the target user into the relaxation state evaluation model generated through training to determine the current relaxation state of the target user, and then according to the target user
  • the user's current relaxation state and the relaxation parameters corresponding to the plurality of music to be recommended determine at least one target recommended music, and the target recommended music is used to play to the target user. Therefore, by wearing a wearable device on the target user, the relaxation state can be evaluated in real time, and this method takes into account the portability of the wearable device and the fitting ability of machine learning technology, and can accurately and effectively provide target users with Music that brings a relaxing effect.
  • steps may be reordered, added or deleted using the various forms of flow shown above.
  • each step described in the present disclosure may be executed in parallel, sequentially, or in a different order, as long as the desired result of the technical solution disclosed in the present disclosure can be achieved, no limitation is imposed herein.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Software Systems (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Artificial Intelligence (AREA)
  • Databases & Information Systems (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Multimedia (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

一种基于可穿戴设备的音乐推荐方法，装置，计算机设备及存储介质，涉及计算机技术领域。本音乐推荐方法包括：获取可穿戴设备采集的目标用户的生理参数（S101）；将目标用户的生理参数，输入经过训练生成的放松状态评估模型中，以确定目标用户当前的放松状态（S102）；根据目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐（S103），其中目标推荐音乐用于播放给目标用户（S104）。

Description

基于可穿戴设备的音乐推荐方法、装置、设备及存储介质
相关申请的交叉引用
本申请基于申请号为202111062501.9、申请日为2021年09月10日的中国专利申请提出，并要求该中国专利申请的优先权，该中国专利申请的全部内容在此引入本申请作为参考。
技术领域
本公开涉及计算机技术领域，具体涉及一种基于可穿戴设备的音乐推荐方法、装置、设备及存储介质。
背景技术
人类的生命当中有将近三分之一的时间是在睡眠，睡眠质量的好坏是影响人类生产生活以及身心健康的重要因素，睡眠问题也越来越受到公共卫生领域的重视。音乐疗法作为一种简单易行的非药物方法就被证明对失眠问题有显著效果。
传统的助眠音乐推荐方法依赖于用户的交互行为或者主观的偏好选择，具有主观不确定性。比如仅仅是按照某个歌单上的歌曲进行循环播放，其播放顺序与助眠效应并没有绝对的联系，因此如何实现精准的个性化助眠音乐推荐是一个重大的技术挑战。
发明内容
本公开提供了一种基于可穿戴设备的音乐推荐方法、装置、设备以及存储介质。
根据本公开的一方面，提供了一种基于可穿戴设备的音乐推荐方法，包括：
获取所述可穿戴设备采集的目标用户的生理参数；
将所述目标用户的生理参数，输入经过训练生成的放松状态评估模型中，以确定所述目标用户当前的放松状态；
根据所述目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐，其中目标推荐音乐用于播放给所述目标用户。
在一些实施例子中，根据所述目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐，包括：
根据所述多个待推荐音乐对应的放松参数，从大到小排序所述多个待推荐音乐；
响应于所述目标用户当前的放松状态为放松，从按照从大到小排序的多个待推荐音乐中按照从前往后的顺序确定第一数量的目标推荐音乐；
响应于所述目标用户当前的放松状态为不放松，从按照从大到小排序的多个待推荐音乐中按照从前往后的顺序确定第二数量的目标推荐音乐；
其中，所述第二数量大于所述第一数量。
在一些实施例中，在将所述目标用户的生理参数，输入经过训练生成的放松状态评估模型之前，还包括：
获取第一训练数据集，其中所述第一训练数据集中包括多个参考用户的生理参数及根据所述多个参考用户的脑电数据标注的放松状态；
将所述参考用户的生理参数输入初始的放松状态评估模型中，以获取预测的放松状态；
根据标注的放松状态与预测的放松状态的差异，对所述初始的放松状态评估模型进行修正，以生成所述经过训练生成的放松状态评估模型。
在一些实施例中，在所述根据所述目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐之前，还包括：
获取所述目标用户的音乐偏好、属性信息及历史睡眠数据；
根据所述目标用户的音乐偏好、属性信息、历史睡眠数据及当前的放松状态，确定所述目标用户当前对应的目标音乐风格；
根据所述多个待推荐音乐所属的音乐风格，从候选音乐库中获取属于所述目标音乐风格的候选音乐作为所述多个待推荐音乐。
在一些实施例中，在所述根据所述目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐之前，还包括：
获取所述多个待推荐音乐对应的音乐特征及所述目标用户的属性信息；
将所述多个待推荐音乐对应的音乐特征及所述目标用户的属性信息输入经过训练生成的放松参数预估模型中，以确定所述多个待推荐音乐对应的放松参数。
在一些实施例中，所述方法，还包括：
获取第二训练数据集，其中所述第二训练数据集中包括根据多个参考用户分别在播放所述多个待推荐音乐时的第一放松曲线及在未播放待推荐音乐时的第二放松曲线；
标注每个所述待推荐音乐对每个所述参考用户的放松参数；
将每个所述待推荐音乐的音乐特征及每个所述参考用户的属性信息，输入初始放松参数预估模型中，以获取每个所述待推荐音乐对每个所述参考用户的预测的放松参数；
根据标注的放松参数与预测的放松参数的差异，对所述初始的放松参数预估模型进行修正，以生成所述经过训练生成的放松参数预估模型。
在一些实施例中，所述目标推荐音乐有多个，所述播放所述至少一个目标推荐音乐，包括：
将所述多个目标推荐音乐进行两两融合，以获取多个融合后的音乐；
根据每个所述融合后的音乐对应的两个目标推荐音乐，确定所述多个目标推荐音乐及融合后的音乐的播放顺序；
基于所述播放顺序，依次播放所述多个目标推荐音乐及融合后的音乐。
在一些实施例中，所述将所述多个目标推荐音乐进行两两融合，以获取多个融合后的音乐，包括：
根据播放顺序，由后至前从每两个目标推荐音乐中的前一个音乐中截取预设时长的第一音乐片段；
根据播放顺序，由前至后从每两个目标推荐音乐中的后一个音乐中截取预设时长的第二音乐片段；
将所述第一音乐片段及所述第二音乐片段进行融合，以获取融合后的音乐。
在一些实施例中，所述将所述第一音乐片段及所述第二音乐片段进行融合，以获取融合后的音乐，包括：
基于第一权重序列和第二权重序列，将所述第一音乐片段及所述第二音乐片段中的各个子片段分别进行融合，以生成融合后的音乐，
其中，所述第一权重序列及所述第二权重序列中分别包含多个权重值，且所述第一权重序列中多个第一权重值逐渐降低，所述第二权重序列中多个第二权重值逐渐增加，且每个第一权重值与对应的第二权重值的和为1。
根据本公开的第二方面，提供了一种基于可穿戴设备的音乐推荐装置，包括：
获取模块，用于获取所述可穿戴设备采集的目标用户的生理参数；
第一确定模块，用于将所述目标用户的生理参数，输入经过训练生成的放松状态评估模型中，以确定所述目标用户当前的放松状态；
第二确定模块，用于根据所述目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐，其中所述目标推荐音乐用于播放给所述目标用户。
根据本公开的第三方面，提供了一种电子设备，包括：
至少一个处理器；以及
与所述至少一个处理器通信连接的存储器；其中，
所述存储器存储有可被所述至少一个处理器执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器能够执行上述一方面实施例所述的方法。
根据本公开的第四方面，提供了一种存储有计算机指令的非瞬时计算机可读存储介质，其上存储有计算机程序，所述计算机指令用于使所述计算机执行上述一方面实施例所述的方法。
根据本公开的第五方面，提供了一种计算机程序产品，包括计算机程序，所述计算机程序在被处理器执行时实现上述一方面实施例所述的方法
本公开实施例中该装置首先获取可穿戴设备采集的目标用户生的理参数，将目标用户的生理参数，输入经过训练生成的放松状态评估模型中，以确定目标用户当前的放松状态，然后根据目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐，目标推荐音乐用于播放给所述目标用户。由此，通过目标用户佩戴可穿戴设备，可以实时的对放松状态进行评估，且该方法兼顾了可穿戴设备的便携性和机器学习技术的拟合能力，能够准确、有效的为目标用户提供可以带来放松效果的音乐。
进一步地，本公开实施例中该装置首先获取第一训练数据集，其中第一训练数据集中包括多个参考用户的生理参数及根据多个参考用户的脑电数据标注的放松状态，，将参考用户的生理参数输入初始的放松状态评估模型中，以获取预测的放松状态，之后根据标注的放松状态与预测的放松状态的差异，对初始的放松状态评估模型进行修正，以 生成经过训练生成的放松状态评估模型，然后获取可穿戴设备采集的目标用户的生理参数，之后将目标用户的生理参数，输入经过训练生成的放松状态评估模型中，以确定目标用户当前的放松状态，再然后根据目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐，目标推荐音乐用于播放给所述目标用户。由此，利用参考用户确定训练数据集并对放松状态评估模型进行修正，可以更准确的确定各个待推荐音乐的放松参数，因而可以更为准确、有效的为目标用户提供可以带来放松效果的音乐。
进一步地，本公开实施例中该装置首先获取可穿戴设备采集的目标用户的生理参数，将目标用户的生理参数，输入经过训练生成的放松状态评估模型中，以确定目标用户当前的放松状态，然后根据目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐，之后将多个目标推荐音乐进行两两融合，以获取多个融合后的音乐，并根据每个融合后的音乐对应的两个目标推荐音乐，确定多个目标推荐音乐及融合后的音乐的播放顺序，最后基于播放顺序，依次播放多个目标推荐音乐及融合后的音乐。由此，通过将目标推荐音乐进行融合后的音乐及对应的目标推荐音乐按一定顺序进行播放，从而不仅可以为不同的目标用户提供个性化的减压放松音乐，避免目标用户对单调曲库产生厌倦，而且使得音乐切换过程无感，能够更准确、更有效的为目标用户提供可以带来放松效果的音乐。
应当理解，本部分所描述的内容并非旨在标识本公开的实施例的关键或重要特征，也不用于限制本公开的范围。本公开的其它特征将通过以下的说明书而变得容易理解。
附图说明
附图用于更好地理解本方案，不构成对本公开的限定。其中：
图1是根据本公开提供的一种基于可穿戴设备的音乐推荐方法的流程示意图；
图2是根据本公开提供的另一种基于可穿戴设备的音乐推荐方法的流程示意图；
图3是根据本公开提供的又一种基于可穿戴设备的音乐推荐方法的流程示意图；
图4为本公开提供的一种基于可穿戴设备的音乐推荐装置的结构框图；
图5为本公开提供的电子设备的结构框图。
具体实施方式
以下结合附图对本公开的示范性实施例做出说明，其中包括本公开实施例的各种细节以助于理解，应当将它们认为仅仅是示范性的。因此，本领域普通技术人员应当认识到，可以对这里描述的实施例做出各种改变和修改，而不会背离本公开的范围和精神。同样，为了清楚和简明，以下的描述中省略了对公知功能和结构的描述。
本公开提出的基于可穿戴设备的音乐推荐方法可由本公开提供的基于可穿戴设备的音乐推荐装置执行，也可以由本公开提供的电子设备执行，其中，电子设备可以包括但不限于台式电脑、平板电脑等终端设备，也可以是服务器，下面以由本公开提供的助 眠音乐的生成装置来执行本公开提供的一种基于可穿戴设备的音乐推荐方法，而不作为对本公开的限定，以下简称为“装置”。
下面参考附图对本公开提供的基于可穿戴设备的音乐推荐方法、装置、计算机设备及存储介质进行详细描述。
图1是根据本公开一实施例的基于可穿戴设备的音乐推荐方法的流程示意图。如图1所示，该基于可穿戴设备的音乐推荐方法可以包括以下步骤S101至步骤S104。
步骤S101，获取可穿戴设备采集的目标用户的生理参数。
可以理解的是，可穿戴设备可以由目标用户穿戴，通过传感器接触或者压靠目标用户的皮肤，以获得目标用户的生理参数。
其中，目标用户的生理参数可以为目标用户的心率、呼吸率、心率变异性、呼吸率变异性、皮肤电阻抗、血压、血氧和睡眠分期等。
可以理解的是，目标用户的生理参数可以作为指标，用于之后对目标用户的放松状态进行评估。举例来说，通常在医学研究中，人体在放松状态下的心率会降低，根据该先验知识，可以认为若人体的心率出现显著降低，则说明此时人体处于放松状态。
步骤S102，将目标用户的生理参数，输入经过训练生成的放松状态评估模型中，以确定目标用户当前的放松状态。
需要说明的是，减压助眠的音乐推荐依赖于对目标用户自身身体放松状态的精准评估。因此该装置在实时地获取目标用户的生理参数后，需要将生理参数数据输入预先训练生成的模型当中，以估计目标用户当前的放松状态。
其中，该装置可以通过放松状态评估模型为目标用户生成对应的放松状态的标签，比如“放松”、“较为放松”、“不放松”、“紧张”。
步骤S103，根据目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐。
其中，目标推荐音乐可以为能够在考虑目标用户自身的特点的同时，辅助目标用户放松助眠的音乐。
在一些实施例中，该装置在根据目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐时，可以有多种方式。
需要说明的是，多个待推荐音乐对应有放松参数，其中，放松参数的大小用于表征待推荐音乐的放松效能。举例来说，若放松参数和待推荐音乐的放松效能成正比，则说明放松参数越大，待推荐音乐的放松效能越高，也即，越有利于目标用户放松、解压、助眠。另外，相同音乐对于不同用户的放松参数也可以是不同的。
举例来说，若待推荐的音乐有M首，对应的放松参数按照从大到小分别为E1、E2、…、EM，那么该装置则可以按照从前往后的顺序从M首音乐中选取一定数量的音乐推荐给目标用户。
比如，若目标用户当前的状态为“放松”，则可以根据放松参数由大到小，推荐前M/4的音乐曲目作为目标推荐音乐给目标用户。或在，若目标用户当前的状态为“不放 松”，则可以从中选取前3M/4的音乐曲目作为目标推荐音乐给目标用户。
步骤S104，播放目标推荐音乐给目标用户。
在一些实施例中，该装置可以通过播放装置播放目标推荐音乐。可以理解的是，在播放目标推荐音乐的同时，该装置还可以利用可穿戴设备实时的采集目标用户的生理参数，从而之后可以对音乐的放松参数进行动态的调整。
本公开实施例中该装置首先获取可穿戴设备采集的目标用户的生理参数，将目标用户的生理参数，输入经过训练生成的放松状态评估模型中，以确定目标用户当前的放松状态，然后根据目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐，目标推荐音乐用于播放给目标用户。由此，通过目标用户佩戴可穿戴设备，可以实时的对放松状态进行评估，且该方法兼顾了可穿戴设备的便携性和机器学习技术的拟合能力，能够准确、有效的为目标用户提供可以带来放松效果的音乐。
图2是根据本公开另一实施例的基于可穿戴设备的音乐推荐方法的流程示意图。
如图2所示，该基于可穿戴设备的音乐推荐方法可以包括以下步骤S201至步骤S207。
步骤S201，获取第一训练数据集，其中第一训练数据集中包括多个参考用户的生理参数及根据多个参考用户的脑电数据标注的放松状态。
需要说明的是，第一训练数据集可以在参考用户中采集。通过参考用户佩戴可穿戴设备，该装置可以实时地对利用可穿戴设备对参考用户的生理参数数据进行采集，比如心率、呼吸频率、皮肤电阻抗、血压、血氧等生理参数，另外，该装置还可以通过脑电采集装置采集各参考用户对应的脑电数据。
在一些实施例中，可以通过将脑电采集装置在人的头皮上检测脑电波信号，并通过相关的设备进行脑电波的收集和处理，从而获取脑电数据。
可以理解的是，由于脑电是可以较为直接和准确衡量人体放松状态的标准，因而本公开可以根据各参考用户的脑电数据，确定各参考用户的标注放松状态。
其中，标注放松状态可以有多种，比如放松、较为放松、略微放松、不放松等。
在一些实施例中，脑电数据可以包括参考用户的脑电波信号，比如α波、β波、θ波、顶尖波、睡眠纺锤波等。通过对脑电波信号特征的分析，该装置可以对参考用户的标注放松状态进行确定。举例来说，若某一参考用户的α波特征的参数为8至13Hz，幅度为20至100μV，则可以确定当前用户的状态为“较为放松”。
需要说明的是，上述举例仅为示意性说明，而不作为对本公开的限定。
步骤S202，将参考用户的生理参数输入初始的放松状态评估模型中，以获取预测的放松状态。
可以理解的是，初始的放松状态评估模型可以为尚未达到可用状态的网络模型，该装置可以通过之后对其网络参数进行修正、训练，以使其达到可用状态。
在一些实施例中，通过将参考用户的生理参数输入初始的放松状态评估模型，以对参考用户的放松状态进行预测，其中，标注放松状态可以有放松、较为放松、略微放松、不放松等多种。
步骤S203，根据标注的放松状态与预测的放松状态的差异，对初始的放松状态评估模型进行修正，以生成经过训练生成的放松状态评估模型。
具体来说，该装置可以将标注放松状态与预测放松状态进行比较，以确定标注放松状态与预测放松状态之间的差异，比如，可以使用梯度下降、随机梯度下降等方式确定出修正梯度，进而以此对初始评估模型进行修正，以生成经过训练生成的放松状态评估模型，本公开修正梯度的方式可以根据实际需要确定。
步骤S204，获取可穿戴设备采集的目标用户的生理参数。
步骤S205，将目标用户的生理参数，输入经过训练生成的放松状态评估模型中，以确定目标用户当前的放松状态。
需要说明的是，步骤S204、S205的具体实现过程可以参照上述实施例，在此不进行赘述。
步骤S206，根据目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐，其中目标推荐音乐用于播放给所述目标用户。
需要说明的是，该装置在根据目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐之前，还需要先获取多个待推荐音乐对应的放松参数。
在一些实施例中，可以将各个待推荐音乐的音乐特征及每个参考用户的属性信息输入预先训练的放松参数预估模型中，以确定各个待推荐音乐的放松参数。为方便理解，下面本公开将对放松参数预估模型的一些实现的生成方式进行说明，在一些实施例中，可以包括以下步骤：
获取第二训练数据集，其中第二训练数据集中包括根据多个参考用户分别在播放所述多个待推荐音乐时的第一放松曲线及在未播放待推荐音乐时的第二放松曲线；
标注每个待推荐音乐对每个参考用户的放松参数；
将每个待推荐音乐的音乐特征及每个参考用户的属性信息，输入初始放松参数预估模型中，以获取每个待推荐音乐对每个参考用户的预测的放松参数；
根据标注的放松参数与预测的放松参数的差异，对初始的放松参数预估模型进行修正，以生成经过训练生成的放松参数预估模型。
其中，第一放松曲线和第二放松曲线都可以表示为时间的函数图像，比如，以时间为横坐标，疲劳度为纵坐标，或者以时间为横坐标，放松程度为纵坐标。需要说明的是，参考用户在播放不同风格的音乐时的放松曲线可以是相同或不同的。
其中，音乐特征可以为音乐的节律、音乐间隔、和声、音调等。需要说明的是，在将音乐特征输入初始的放松参数预估模型之前，可以对各个音乐的音乐特征进行量化。其中，参考用户的属性信息可以为参考用户的年龄、性别、体重、音乐偏好等信息。
其中，放松参数也即音乐对应的放松效能，不同风格种类的音乐的放松参数不同。另外，相同音乐对于不同的目标用户来说，放松参数也不同。
需要说明的是，由于第二放松曲线为未播放待推荐音乐状态下的放松曲线，该装置 可以通过计算第一放松曲线和第二放松曲线在时间上的积分之差，标注每个待推荐音乐对每个参考用户的放松参数。
其中，初始的放松参数预估模型可以为尚未达到可用状态的网络模型，该装置可以通过之后对其网络参数进行修正、训练，以使其达到可用状态。在一些实施例中，通过将每个待推荐音乐的音乐特征及每个参考用户的属性信息输入初始的放松参数预估模型，以获取每个待推荐音乐对每个参考用户的预测的放松参数。
在一些实施例中，该装置可以将标注的放松参数与预测的放松参数进行比较，以确定标注放松参数与预测放松参数之间的差异，比如，可以使用梯度下降、随机梯度下降等方式确定出修正梯度，进而以此对初始放松参数预估模型进行修正，以生成经过训练生成的放松参数预估模型，本公开修正梯度的方式可以根据实际需要确定。
进一步地，在确定放松参数预估模型之后，该装置可以通过输入每个待推荐音乐的音乐特征及每个目标用户的属性信息，以确定多个待推荐音乐对应的放松参数。在一些实施例中，还可以输入之前时间窗口的目标用户的放松状态，比如若以4小时为一个时间窗口，则可以输入当前时间4小时之前目标用户的放松状态。
之后，该装置可以根据目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐。
步骤S207，播放目标推荐音乐给目标用户。
需要说明的是，步骤S207的具体实现过程可以参照上述实施例，在此不进行赘述。
本公开实施例中该装置首先获取第一训练数据集，其中第一训练数据集中包括多个参考用户的生理参数及根据多个参考用户的脑电数据标注的放松状态，将参考用户的生理参数输入初始的放松状态评估模型中，以获取预测的放松状态，之后根据标注的放松状态与预测的放松状态的差异，对初始的放松状态评估模型进行修正，以生成经过训练生成的放松状态评估模型，然后获取可穿戴设备采集的目标用户的生理参数，之后将目标用户的生理参数，输入经过训练生成的放松状态评估模型中，以确定目标用户当前的放松状态，再然后根据目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐，目标推荐音乐用于播放给所述目标用户。由此，利用参考用户确定训练数据集并对放松状态评估模型进行修正，可以更准确的确定各个待推荐音乐的放松参数，因而可以更为准确、有效的为目标用户提供可以带来放松效果的音乐。
图3是根据本公开又一实施例的基于可穿戴设备的音乐推荐方法的流程示意图。
如图3所示，该基于可穿戴设备的音乐推荐方法可以包括以下步骤S301至步骤S306。
步骤S301，获取可穿戴设备采集的目标用户的生理参数。
步骤S302，将目标用户生理参数，输入经过训练生成的放松状态评估模型中，以确定目标用户当前的放松状态。
需要说明的是，步骤S301、S302的具体实现过程可以参照上述任一实施例，本公开在此不进行赘述。
步骤S303，根据目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确 定至少一个目标推荐音乐。
需要说明的是，待推荐音乐可以为预置于候选音乐库中的候选音乐，可以有多种不同风格，在确定至少一个目标推荐音乐之前，该装置可以获取部分待推荐音乐，下面本公开将对获取待推荐音乐的一些实现的方式进行说明。
在一些实施例中，该装置可以首先获取目标用户的音乐偏好、属性信息及历史睡眠数据，然后根据目标用户的音乐偏好、属性信息、历史睡眠数据及当前的放松状态，确定目标用户当前对应的目标音乐风格，之后根据所述多个待推荐音乐所属的音乐风格，从候选音乐库中获取属于目标音乐风格的候选音乐作为多个待推荐音乐。
可以理解的是，由于用户的多样性，不同用户所适用的助眠音乐也千变万化，为实现精准的个性化音乐推荐，可以首先获取用户的音乐偏好，比如，国风音乐、纯音乐、白噪声、古典乐、美声音乐等，用户的属性信息，比如，年龄、性别、身高、体重等，以及历史睡眠数据，比如当前时间的前一天此时用户的放松状态。
需要说明的是，上述对用户的音乐偏好、属性信息及历史睡眠数据的举例仅为示意性说明。
进而，该装置可以确定目标用户当前对应的目标音乐风格，并在候选音乐库中获取当前用户所需求的属于目标音乐风格的候选音乐作为多个待推荐音乐。其中，待推荐音乐可以为多个音乐。
步骤S304，将多个目标推荐音乐进行两两融合，以获取多个融合后的音乐。
需要说明的是，在将多个目标推荐音乐进行两两融合时，可以有多种方式。
在一些实施例中，可以根据多个目标推荐音乐的播放顺序，将每相邻的两个目标推荐音乐进行加权融合，以获取多个融合后的音乐。
在一些实施例中，该装置还可以首先根据播放顺序，由后至前从每两个目标推荐音乐中的前一个音乐中截取预设时长的第一音乐片段，然后根据播放顺序，由前至后从每两个目标推荐音乐中的后一个音乐中截取预设时长的第二音乐片段。
可以理解的是，若目标推荐音乐为多个，在确定多个目标推荐音乐之后，该装置可以将目标推荐音乐按照放松参数从大到小的方式，确定播放顺序，然后再将每两个相邻的目标推荐音乐进行融合。
举例来说，若按照放松参数，确定的至少一个目标推荐音乐的播放顺序为A、B、C，该装置可以将目标推荐音乐两两组合为(A,B)和(B,C)。其中，对于第一组(A,B)，该装置可以由后至前，从音乐A中选取预设时长(比如为20s)的第一音乐片段A1，并由前至后从音乐B中选取20s的第二音乐片段B1。
进一步地，将第一音乐片段及第二音乐片段进行融合，以获取多个融合后的音乐。需要理解的是，在将第一音乐片段及第二音乐片段进行融合时，可以有多种方式。
在一些实施例中，可以将第一音乐片段及第二音乐片段中的各个子片段分别相加，从而使得融合后的音乐中包含了两个目标推荐音乐的元素。
在一些实施例中，该装置还可以基于第一权重序列和第二权重序列，将第一音乐片 段及第二音乐片段中的各个子片段分别进行融合，以生成融合后的音乐。其中，第一权重序列及第二权重序列中分别包含多个权重值，且第一权重序列中多个第一权重值逐渐降低，第二权重序列中多个第二权重值逐渐增加，且每个第一权重值与对应的第二权重值的和为1。
其中，第一权重序列为第一音乐片段的各个子片段对应的权重序列，第二权重序列为第二音乐片段的各个子片段对应权重序列。
举例来说，若第一音乐片段为A1，其中，A1包括的子片段有a1、a2、a3,第二音乐片段为B1，其中，B1包括的子片段有b1、b2、b3。第一权重序列a1、a2、a3对应的权重分别为0.8、0.5、0.3，第二权重序列b1、b2、b3对应的权重分别为0.2、0.5、0.7。
通过上述融合方式，可以使得生成的融合音乐中，包含的第一音乐片段A1的比重越来越少，而第二音乐片段B1所占的比重越来越大。
需要说明的是，上述示例仅为本公开的示意性说明。
步骤S305，根据每个融合后的音乐对应的两个目标推荐音乐，确定多个目标推荐音乐及融合后的音乐的播放顺序。
本公开中，为了避免不同目标推荐音乐在顺序播放过程中，带来的音乐切换，造成感官突变，对用户造成干扰。通过将多个目标推荐音乐进行两两融合后，生成的融合音乐中即可包含两个目标推荐音乐的元素。从而在播放时，可以那个融合后的音乐设置在对应的两个目标推荐音乐之间播放，从而可以使得音乐的切换过程无感，改善了用户体验。
举例来说，对于播放顺序为A、B、C的目标推荐音乐，该装置可以将A和B融合后生成的音乐，设置在A和B之间播放，而将B和C融合的音乐设置在B和C之间播放，以此类推。
由此，可以提高目标推荐音乐之间的相似度，避免给用户带来音乐切换的突兀感，且可以更加实时地生成更多放松歌曲丰富曲库音乐。
步骤S306，基于播放顺序，依次播放多个目标推荐音乐及融合后的音乐。
在一些实施例中，该装置可以利用播放装置将多个目标推荐音乐及融合后的音乐按照播放顺序进行播放。另外，还可以由可穿戴设备对目标用户的生理参数进行实时采集，由此，能够根据目标用户的实时放松状态调整推荐音乐。
本公开实施例中该装置首先获取可穿戴设备采集的目标用户的生理参数，将目标用户的生理参数，输入经过训练生成的放松状态评估模型中，以确定目标用户当前的放松状态，然后根据目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐，之后将多个目标推荐音乐进行两两融合，以获取多个融合后的音乐，并根据每个融合后的音乐对应的两个目标推荐音乐，确定多个目标推荐音乐及融合后的音乐的播放顺序，最后基于播放顺序，依次播放多个目标推荐音乐及融合后的音乐。由此，通过将目标推荐音乐进行融合后的音乐及对应的目标推荐音乐按一定顺序进行播放，从而不仅可以为不同的目标用户提供个性化的减压放松音乐，避免目标用户对单调曲库 产生厌倦，而且使得音乐切换过程无感，能够更准确、更有效的为目标用户提供可以带来放松效果的音乐。
为了实现上述实施例，本公开实施例还提出一种基于可穿戴设备的音乐推荐装置。图4为本公开实施例提供的一种基于可穿戴设备的音乐推荐装置的结构框图。
如图4所示，该基于可穿戴设备的音乐推荐装置包括：获取模块410、第一确定模块420、第二确定模块430及播放模块440。
获取模块410，用于获取所述可穿戴设备采集的目标用户的生理参数；
第一确定模块420，用于将所述目标用户生理参数，输入经过训练生成的放松状态评估模型中，以确定所述目标用户当前的放松状态；
第二确定模块430，用于根据所述目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐；
播放模块440，用于播放所述目标推荐音乐给目标用户。
在一些实施例中，所述第一确定模块420，还用于：
获取第一训练数据集，其中所述第一训练数据集中包括多个参考用户的生理参数及根据多个参考用户的脑电数据标注的放松状态；
将所述参考用户的生理参数输入初始的放松状态评估模型中，以获取的预测放松状态；
根据标注的放松状态与预测的放松状态的差异，对所述初始的放松状态评估模型进行修正，以生成所述经过训练生成的放松状态评估模型。
在一些实施例中，所述第二确定模块430，还用于：
获取所述目标用户的音乐偏好、属性信息及历史睡眠数据；
根据所述目标用户的音乐偏好、属性信息、历史睡眠数据及当前的放松状态，确定所述目标用户当前对应的目标音乐风格；
根据所述多个待推荐音乐所属的音乐风格，从候选音乐库中获取属于所述目标音乐风格的候选音乐作为所述多个待推荐音乐。
在一些实施例中，所述第二确定模块430，还用于：
获取多个待推荐音乐对应的音乐特征及所述目标用户的属性信息；
将所述多个待推荐音乐对应的音乐特征及所述目标用户的属性信息输入经过训练生成的放松参数预估模型中，以确定所述多个待推荐音乐对应的放松参数。
在一些实施例中，所述装置，还包括：
第二获取模块，用于获取第二训练数据集，其中所述第二训练数据集中包括根据多个参考用户分别在播放多个待推荐音乐时的第一放松曲线及在未播放待推荐音乐时的第二放松曲线；
第三确定模块，用于标注每个所述待推荐音乐对每个所述参考用户的放松参数；
第三获取模块，用于将每个所述待推荐音乐的音乐特征及每个所述参考用户的属性信息，输入初始放松参数预估模型中，以获取每个所述待推荐音乐对每个所述参考用户 的预测的放松参数；
修正模块，用于根据标注的放松参数与预测的放松参数的差异，对所述初始的放松参数预估模型进行修正，以生成所述经过训练生成的放松参数预估模型。
在一些实施例中，所述播放模块，包括：
融合单元，用于将所述多个目标推荐音乐进行两两融合，以获取多个融合后的音乐；
确定单元，用于根据每个所述融合后的音乐对应的两个目标推荐音乐，确定所述多个目标推荐音乐及融合后的音乐的播放顺序；
播放单元，用于基于所述播放顺序，依次播放所述多个目标推荐音乐及融合后的音乐。
在一些实施例中，所述融合单元，包括：
第一截取子单元，用于根据播放顺序，由后至前从每两个目标推荐音乐中的前一个音乐中截取预设时长的第一音乐片段；
第二截取子单元，用于根据播放顺序，由前至后从每两个目标推荐音乐中的后一个音乐中截取预设时长的第二音乐片段；
融合子单元，用于将所述第一音乐片段及所述第二音乐片段进行融合，以获取融合后的音乐。
在一些实施例中，所述融合子单元，具体用于：
基于第一权重序列和第二权重序列，将所述第一音乐片段及所述第二音乐片段中的各个子片段分别进行融合，以生成融合后的音乐，
其中，所述第一权重序列及所述第二权重序列中分别包含多个权重值，且所述第一权重序列中多个第一权重值逐渐降低，所述第二权重序列中多个第二权重值逐渐增加，且每个第一权重值与对应的第二权重值的和为1。
本公开实施例中该装置首先获取可穿戴设备采集的目标用户的生理参数，将目标用户生理参数，输入经过训练生成的放松状态评估模型中，以确定目标用户当前的放松状态，然后根据目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐，目标推荐音乐用于播放给所述目标用户。由此，通过目标用户佩戴可穿戴设备，可以实时的对放松状态进行评估，且该方法兼顾了可穿戴设备的便携性和机器学习技术的拟合能力，能够准确、有效的为目标用户提供可以带来放松效果的音乐。
根据本公开的实施例，本公开还提供了一种电子设备、一种可读存储介质、一种计算机程序产品。
图5示出了可以用来实施本公开的实施例的示例电子设备500的示意性框图。电子设备旨在表示各种形式的数字计算机，诸如，膝上型计算机、台式计算机、工作台、个人数字助理、服务器、刀片式服务器、大型计算机、和其它适合的计算机。电子设备还可以表示各种形式的移动装置，诸如，个人数字处理、蜂窝电话、智能电话、可穿戴设备和其它类似的计算装置。
在一些实施例中，上述电子设备可以是可穿戴设备，可穿戴设备用于采集目标用户 的生理参数，并且可穿戴设备可以用于根据本公开的音乐推荐方法确定目标推荐音乐，并将目标推荐音乐在可穿戴设备上直接播放给目标用户。
在一些实施例中，上述电子设备可以是移动终端，可穿戴设备采集目标用户的生理参数，并且可穿戴设备将采集的生理参数传输给移动终端(移动终端的用户通常与可穿戴设备的用户是相同的人，在某些情况下移动终端的用户与可穿戴设备的用户不是相同的人)，并由移动终端上运行的应用程序处理。其中，移动终端与可穿戴设备无线通信连接。移动终端上运行的应用程序可以用于根据本公开的音乐推荐方法确定目标推荐音乐，并将目标推荐音乐播放给目标用户。
在一些实施例中，电子设备可以服务器，可穿戴设备用于采集目标用户的生理参数，并且可穿戴设备将采集的生理参数传输给移动终端(移动终端的用户通常与可穿戴设备的用户是相同的人，在某些情况下移动终端的用户与可穿戴设备的用户不是相同的人)，移动终端上运行有应用程序，移动终端将接收到的生理参数传输给远程服务器(例如，云端服务器)，并由远程服务器上运行的程序处理。其中，移动终端与可穿戴设备无线通信连接，远程服务器与移动终端无线通信连接。远程服务器可以用于根据本公开的音乐推荐方法确定目标推荐音乐，并将目标推荐音乐推送给移动终端，以使目标推荐音乐在移动终端上播放给目标用户。
本文所示的部件、它们的连接和关系、以及它们的功能仅仅作为示例，并且不意在限制本文中描述的和/或者要求的本公开的实现。
如图5所示，设备500包括计算单元501，其可以根据存储在只读存储器(ROM)502中的计算机程序或者从存储单元508加载到随机访问存储器(RAM)503中的计算机程序，来执行各种适当的动作和处理。在RAM 503中，还可存储设备500操作所需的各种程序和数据。计算单元501、ROM 502以及RAM 503通过总线504彼此相连。输入/输出(I/O)接口505也连接至总线504。
设备500中的多个部件连接至I/O接口505，包括：输入单元506，例如键盘、鼠标等；输出单元507，例如各种类型的显示器、扬声器等；存储单元508，例如磁盘、光盘等；以及通信单元509，例如网卡、调制解调器、无线通信收发机等。通信单元509允许设备500通过诸如因特网的计算机网络和/或各种电信网络与其他设备交换信息/数据。
计算单元501可以是各种具有处理和计算能力的通用和/或专用处理组件。计算单元501的一些示例包括但不限于中央处理单元(CPU)、图形处理单元(GPU)、各种专用的人工智能(AI)计算芯片、各种运行机器学习模型算法的计算单元、数字信号处理器(DSP)、以及任何适当的处理器、控制器、微控制器等。计算单元501执行上文所描述的各个方法和处理，例如基于可穿戴设备的音乐推荐方法。例如，在一些实施例中，基于可穿戴设备的音乐推荐方法可被实现为计算机软件程序，其被有形地包含于机器可读介质，例如存储单元508。在一些实施例中，计算机程序的部分或者全部可以经由ROM502和/或通信单元509而被载入和/或安装到设备500上。当计算机程序加载到RAM 503 并由计算单元501执行时，可以执行上文描述的基于可穿戴设备的音乐推荐方法的一个或多个步骤。备选地，在其他实施例中，计算单元501可以通过其他任何适当的方式(例如，借助于固件)而被配置为执行基于可穿戴设备的音乐推荐方法。
本文中以上描述的系统和技术的各种实施方式可以在数字电子电路系统、集成电路系统、场可编程门阵列(FPGA)、专用集成电路(ASIC)、专用标准产品(ASSP)、芯片上系统的系统(SOC)、负载可编程逻辑设备(CPLD)、计算机硬件、固件、软件、和/或它们的组合中实现。这些各种实施方式可以包括：实施在一个或者多个计算机程序中，该一个或者多个计算机程序可在包括至少一个可编程处理器的可编程系统上执行和/或解释，该可编程处理器可以是专用或者通用可编程处理器，可以从存储系统、至少一个输入装置、和至少一个输出装置接收数据和指令，并且将数据和指令传输至该存储系统、该至少一个输入装置、和该至少一个输出装置。
用于实施本公开的方法的程序代码可以采用一个或多个编程语言的任何组合来编写。这些程序代码可以提供给通用计算机、专用计算机或其他可编程数据处理装置的处理器或控制器，使得程序代码当由处理器或控制器执行时使流程图和/或框图中所规定的功能/操作被实施。程序代码可以完全在机器上执行、部分地在机器上执行，作为独立软件包部分地在机器上执行且部分地在远程机器上执行或完全在远程机器或服务器上执行。
在本公开的上下文中，机器可读介质可以是有形的介质，其可以包含或存储以供指令执行系统、装置或设备使用或与指令执行系统、装置或设备结合地使用的程序。机器可读介质可以是机器可读信号介质或机器可读储存介质。机器可读介质可以包括但不限于电子的、磁性的、光学的、电磁的、红外的、或半导体系统、装置或设备，或者上述内容的任何合适组合。机器可读存储介质的更具体示例会包括基于一个或多个线的电气连接、便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦除可编程只读存储器(EPROM或快闪存储器)、光纤、便捷式紧凑盘只读存储器(CD-ROM)、光学储存设备、磁储存设备、或上述内容的任何合适组合。
为了提供与目标用户的交互，可以在计算机上实施此处描述的系统和技术，该计算机具有：用于向目标用户显示信息的显示装置(例如，CRT(阴极射线管)或者LCD(液晶显示器)监视器)；以及键盘和指向装置(例如，鼠标或者轨迹球)，目标用户可以通过该键盘和该指向装置来将输入提供给计算机。其它种类的装置还可以用于提供与目标用户的交互；例如，提供给目标用户的反馈可以是任何形式的传感反馈(例如，视觉反馈、听觉反馈、或者触觉反馈)；并且可以用任何形式(包括声输入、语音输入或者、触觉输入)来接收来自目标用户的输入。
可以将此处描述的系统和技术实施在包括后台部件的计算系统(例如，作为数据服务器)、或者包括中间件部件的计算系统(例如，应用服务器)、或者包括前端部件的计算系统(例如，具有图形用户界面或者网络浏览器的用户计算机，目标用户可以通过该图形用户界面或者该网络浏览器来与此处描述的系统和技术的实施方式交互)、或者 包括这种后台部件、中间件部件、或者前端部件的任何组合的计算系统中。可以通过任何形式或者介质的数字数据通信(例如，通信网络)来将系统的部件相互连接。通信网络的示例包括：局域网(LAN)、广域网(WAN)、互联网和区块链网络。
计算机系统可以包括客户端和服务器。客户端和服务器一般远离彼此并且通常通过通信网络进行交互。通过在相应的计算机上运行并且彼此具有客户端-服务器关系的计算机程序来产生客户端和服务器的关系。服务器可以是云服务器，又称为云计算服务器或云主机，是云计算服务体系中的一项主机产品，以解决了传统物理主机与VPS服务("Virtual Private Server"，或简称"VPS")中，存在的管理难度大，业务扩展性弱的缺陷。服务器也可以为分布式系统的服务器，或者是结合了区块链的服务器。
本公开实施例中该装置首先获取可穿戴设备采集的目标用户的生理参数，将目标用户的生理参数，输入经过训练生成的放松状态评估模型中，以确定目标用户当前的放松状态，然后根据目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐，目标推荐音乐用于播放给所述目标用户。由此，通过目标用户佩戴可穿戴设备，可以实时的对放松状态进行评估，且该方法兼顾了可穿戴设备的便携性和机器学习技术的拟合能力，能够准确、有效的为目标用户提供可以带来放松效果的音乐。
应该理解，可以使用上面所示的各种形式的流程，重新排序、增加或删除步骤。例如，本发公开中记载的各步骤可以并行地执行也可以顺序地执行也可以不同的次序执行，只要能够实现本公开公开的技术方案所期望的结果，本文在此不进行限制。
上述具体实施方式，并不构成对本公开保护范围的限制。本领域技术人员应该明白的是，根据设计要求和其他因素，可以进行各种修改、组合、子组合和替代。任何在本公开的精神和原则之内所作的修改、等同替换和改进等，均应包含在本公开保护范围之内。

Claims (15)

  1. 一种基于可穿戴设备的音乐推荐方法，其特征在于，包括：
    获取所述可穿戴设备采集的目标用户的生理参数；
    将所述目标用户的生理参数，输入经过训练生成的放松状态评估模型中，以确定所述目标用户当前的放松状态；
    根据所述目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐，其中所述目标推荐音乐用于播放给所述目标用户。
  2. 如权利要求1所述的方法，其特征在于，根据所述目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐，包括：
    根据所述多个待推荐音乐对应的放松参数，从大到小排序所述多个待推荐音乐；
    响应于所述目标用户当前的放松状态为放松，从排序的多个待推荐音乐中按照从前往后的顺序确定第一数量的目标推荐音乐；
    响应于所述目标用户当前的放松状态为不放松，从排序的多个待推荐音乐中按照从前往后的顺序确定第二数量的目标推荐音乐；
    其中，所述第二数量大于所述第一数量。
  3. 如权利要求1所述的方法，其特征在于，在所述将所述目标用户的生理参数，输入经过训练生成的放松状态评估模型之前，还包括：
    获取第一训练数据集，其中所述第一训练数据集中包括多个参考用户的生理参数及根据所述多个参考用户的脑电数据标注的放松状态；
    将所述参考用户的生理参数输入初始的放松状态评估模型中，以获取预测的放松状态；
    根据标注的放松状态与预测的放松状态的差异，对所述初始的放松状态评估模型进行修正，以生成所述经过训练生成的放松状态评估模型。
  4. 如权利要求1所述的方法，其特征在于，在所述根据所述目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐之前，还包括：
    获取所述目标用户的音乐偏好、属性信息及历史睡眠数据；
    根据所述目标用户的音乐偏好、属性信息、历史睡眠数据及当前的放松状态，确定所述目标用户当前对应的目标音乐风格；
    根据所述多个待推荐音乐所属的音乐风格，从候选音乐库中获取属于所述目标音乐风格的候选音乐作为所述多个待推荐音乐。
  5. 如权利要求1所述的方法，其特征在于，在所述根据所述目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐之前，还包括：
    获取所述多个待推荐音乐对应的音乐特征及所述目标用户的属性信息；
    将所述多个待推荐音乐对应的音乐特征及所述目标用户的属性信息输入经过训练生成的放松参数预估模型中，以确定所述多个待推荐音乐对应的放松参数。
  6. 如权利要求5所述的方法，其特征在于，还包括：
    获取第二训练数据集，其中所述第二训练数据集中包括根据多个参考用户分别在播放所述多个待推荐音乐时的第一放松曲线及在未播放待推荐音乐时的第二放松曲线；标注每个所述待推荐音乐对每个所述参考用户的放松参数；
    将每个所述待推荐音乐的音乐特征及每个所述参考用户的属性信息，输入初始放松参数预估模型中，以获取每个所述待推荐音乐对每个所述参考用户的预测的放松参数；
    根据标注的放松参数与预测的放松参数的差异，对所述初始的放松参数预估模型进行修正，以生成所述经过训练生成的放松参数预估模型。
  7. 如权利要求1至6任一所述的方法，其特征在于，所述目标推荐音乐有多个，所述播放所述至少一个目标推荐音乐，包括：
    将所述多个目标推荐音乐进行两两融合，以获取多个融合后的音乐；
    根据每个所述融合后的音乐对应的两个目标推荐音乐，确定所述多个目标推荐音乐及融合后的音乐的播放顺序；
    基于所述播放顺序，依次播放所述多个目标推荐音乐及融合后的音乐。
  8. 如权利要求7所述的方法，其特征在于，所述将所述多个目标推荐音乐进行两两融合，以获取多个融合后的音乐，包括：
    根据播放顺序，由后至前从每两个目标推荐音乐中的前一个音乐中截取预设时长的第一音乐片段；
    根据播放顺序，由前至后从每两个目标推荐音乐中的后一个音乐中截取预设时长的第二音乐片段；
    将所述第一音乐片段及所述第二音乐片段进行融合，以获取融合后的音乐。
  9. 如权利要求8所述的方法，其特征在于，所述将所述第一音乐片段及所述第二音乐片段进行融合，以获取融合后的音乐，包括：
    基于第一权重序列和第二权重序列，将所述第一音乐片段及所述第二音乐片段中的各个子片段分别进行融合，以生成融合后的音乐，
    其中，所述第一权重序列及所述第二权重序列中分别包含多个权重值，且所述第一权重序列中多个第一权重值逐渐降低，所述第二权重序列中多个第二权重值逐渐增加，且每个第一权重值与对应的第二权重值的和为1。
  10. 一种基于可穿戴设备的音乐推荐装置，其特征在于，包括：
    获取模块，用于获取所述可穿戴设备采集的目标用户的生理参数；
    第一确定模块，用于将所述目标用户的生理参数，输入经过训练生成的放松状态评估模型中，以确定所述目标用户当前的放松状态；
    第二确定模块，用于根据所述目标用户当前的放松状态及多个待推荐音乐对应的放松参数，确定至少一个目标推荐音乐，其中所述目标推荐音乐用于播放给所述目标用户。
  11. 一种电子设备，包括：
    至少一个处理器；以及
    与所述至少一个处理器通信连接的存储器；其中，
    所述存储器存储有可被所述至少一个处理器执行的指令，所述指令被所述至少一个处理器执行，以使所述至少一个处理器能够执行权利要求1至9中任一项所述的方法。
  12. 如权利要求11所述的电子设备，其中，所述电子设备为可穿戴设备、与所述可穿戴设备无线通信的移动终端或者与所述移动终端无线通信的远程服务器。
  13. 如权利要求12所述的电子设备，其中，所述电子设备与所述可穿戴设备之间近距离无线通信。
  14. 一种存储有计算机指令的非瞬时计算机可读存储介质，其中，所述计算机指令用于使所述计算机执行根据权利要求1至9中任一项所述的方法。
  15. 一种计算机程序产品，包括计算机程序，所述计算机程序在被处理器执行时实现根据权利要求1至9中任一项所述的方法。
PCT/CN2022/092080 2021-09-10 2022-05-10 基于可穿戴设备的音乐推荐方法、装置、设备及存储介质 WO2023035647A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202111062501.9 2021-09-10
CN202111062501.9A CN115795085A (zh) 2021-09-10 2021-09-10 基于可穿戴设备的音乐推荐方法、装置、设备及存储介质

Publications (1)

Publication Number Publication Date
WO2023035647A1 true WO2023035647A1 (zh) 2023-03-16

Family

ID=85417284

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/092080 WO2023035647A1 (zh) 2021-09-10 2022-05-10 基于可穿戴设备的音乐推荐方法、装置、设备及存储介质

Country Status (2)

Country Link
CN (1) CN115795085A (zh)
WO (1) WO2023035647A1 (zh)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101120412A (zh) * 2005-02-14 2008-02-06 皇家飞利浦电子股份有限公司 一种用于将第一音频数据和第二音频数据混合的系统和方法、一种程序单元以及一种计算机可读介质
US20150297109A1 (en) * 2014-04-22 2015-10-22 Interaxon Inc. System and method for associating music with brain-state data
CN107515925A (zh) * 2017-08-24 2017-12-26 北京小米移动软件有限公司 音乐播放方法及装置
CN107997751A (zh) * 2018-01-16 2018-05-08 华南理工大学 一种基于生物反馈的智能耳机系统
CN108543193A (zh) * 2018-03-14 2018-09-18 天使智心（北京）科技有限公司 一种用户状态干预方法及装置
US20180314959A1 (en) * 2017-05-01 2018-11-01 International Business Machines Corporation Cognitive music selection system and method
CN112121283A (zh) * 2020-09-08 2020-12-25 杭州趣安科技有限公司 一种自适应放松辅助系统、方法、设备以及存储介质
CN112435641A (zh) * 2020-11-09 2021-03-02 腾讯科技（深圳）有限公司 音频处理方法、装置、计算机设备及存储介质

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101120412A (zh) * 2005-02-14 2008-02-06 皇家飞利浦电子股份有限公司 一种用于将第一音频数据和第二音频数据混合的系统和方法、一种程序单元以及一种计算机可读介质
US20150297109A1 (en) * 2014-04-22 2015-10-22 Interaxon Inc. System and method for associating music with brain-state data
US20180314959A1 (en) * 2017-05-01 2018-11-01 International Business Machines Corporation Cognitive music selection system and method
CN107515925A (zh) * 2017-08-24 2017-12-26 北京小米移动软件有限公司 音乐播放方法及装置
CN107997751A (zh) * 2018-01-16 2018-05-08 华南理工大学 一种基于生物反馈的智能耳机系统
CN108543193A (zh) * 2018-03-14 2018-09-18 天使智心（北京）科技有限公司 一种用户状态干预方法及装置
CN112121283A (zh) * 2020-09-08 2020-12-25 杭州趣安科技有限公司 一种自适应放松辅助系统、方法、设备以及存储介质
CN112435641A (zh) * 2020-11-09 2021-03-02 腾讯科技（深圳）有限公司 音频处理方法、装置、计算机设备及存储介质

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
AYATA DEGER; YASLAN YUSUF; KAMASAK MUSTAFA E.: "Emotion Based Music Recommendation System Using Wearable Physiological Sensors", IEEE TRANSACTIONS ON CONSUMER ELECTRONICS, IEEE SERVICE CENTER, NEW YORK, NY, US, vol. 64, no. 2, 1 May 2018 (2018-05-01), NEW YORK, NY, US , pages 196 - 203, XP011686795, ISSN: 0098-3063, DOI: 10.1109/TCE.2018.2844736 *
PENG CHENG, CHANG XIANGMAO; QIU YUAN: "Heart rate variability analysis based sleep music recommendation system", JOURNAL OF COMPUTER APPLICATIONS, JISUANJI YINGYONG, CN, vol. 40, no. 5, 10 May 2020 (2020-05-10), CN , pages 1539 - 1544, XP093045688, ISSN: 1001-9081, DOI: 10.11772/j.issn.1001-9081.2019111969 *

Also Published As

Publication number Publication date
CN115795085A (zh) 2023-03-14

Similar Documents

Publication Publication Date Title
US20200275848A1 (en) Virtual reality guided meditation with biofeedback
US11164596B2 (en) Sensor assisted evaluation of health and rehabilitation
US10172517B2 (en) Image-analysis for assessing heart failure
Wen et al. Consumers’ perceived attitudes to wearable devices in health monitoring in China: A survey study
US20220117515A1 (en) Advanced Consumer Application for Mobile Assessment of Functional Capacity and Falls Risk
CN107106028B (zh) 用于心肺睡眠阶段分类的系统和方法
US11723568B2 (en) Mental state monitoring system
EP4052262A1 (en) Method and system for an interface to provide activity recommendations
WO2018090304A1 (zh) 一种精神压力评测方法和装置
US11673025B2 (en) Workout recommendation engine
Liu et al. Beatquency domain and machine learning improve prediction of cardiovascular death after acute coronary syndrome
Raij et al. mstress: Supporting continuous collection of objective and subjective measures of psychosocial stress on mobile devices
JP7266548B2 (ja) 訓練性ストレス情報を用いて体調を推定する装置、プログラム及び方法
Hardy et al. Associations between voice and gestural characteristics of transgender women and self-rated femininity, satisfaction, and quality of life
EP3903317A1 (en) System and method for optimizing sleep-related parameters for computing a sleep score
CN108697363A (zh) 用于检测心脏变时功能不全的装置和方法
US20220134048A1 (en) Systems and methods for virtual-reality enhanced quantitative meditation
WO2023035647A1 (zh) 基于可穿戴设备的音乐推荐方法、装置、设备及存储介质
Van et al. A Personalized Adaptive Algorithm for Sleep Quality Prediction using Physiological and Environmental Sensing Data
WO2023179765A1 (zh) 多媒体推荐方法及其装置
Nurwahidin et al. Stress detection due to lack of rest using artificial neural network (ann)
WO2022208873A1 (ja) ストレス推定装置、ストレス推定方法及び記憶媒体
US20240058568A1 (en) Methods And Apparatuses For Generating Sleeping Aid Audio
US20240049981A1 (en) Systems and methods for estimation of forced vital capacity using speech acoustics
US20240008813A1 (en) Smart wearable device and method for estimating traditional medicine system parameters

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22866125

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE