Evaluation method and device of AIGC music curing system

Abstract

The invention discloses an evaluation method and device for an AIGC music curing system, and the method comprises the steps: (1) collecting a physiological signal of a listener in real time in the process of carrying out music curing through the AIGC music curing system, and calculating the synchronization rate of a music signal and the physiological signal; (2) constructing a melody prediction model based on an LSTM network, outputting probability distribution of a next note after inputting a historical note sequence, and further calculating a melody predictability score; (3) obtaining a subjective index and an objective index in the music curing process, standardizing the subjective index and the objective index, calculating the current tense, and calculating the consistency of the subjective index and the objective index; and (4) constructing a dynamic weighted total evaluation function based on the synchronization rate, the melody predictability score, the current tense and the consistency, and evaluating the AIGC music curing system according to the result of the total evaluation function. According to the invention, the evaluation accuracy can be improved, and the adaptability of the system and the individuation level of treatment can be improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of music therapy, and in particular relates to an evaluation method and device for an AIGC music healing system. Background Art

[0002] Mental health issues have become increasingly prominent in recent years. Traditional mental illness treatments, based on a pathological medical model, have significant limitations: drug interventions are prone to side effects, while non-drug interventions (such as psychological counseling) are costly and rely on physician experience, making personalized, precise treatment difficult to achieve.

[0003] AIGC-based music therapy has become an emerging treatment due to its customizability, low cost, and ease of adoption. AIGC technology generates adaptive music in real time based on a patient's emotional state, modulating neural activity through sound wave frequency and rhythm to alleviate symptoms such as anxiety and depression.

[0004] For example, a Chinese patent document with publication number CN115331781A discloses a depression treatment system based on electrocardiogram signals and music. The system uses a biosensor to collect the human heart rate, tracks the individual's sympathetic and parasympathetic nerve activity patterns in real time based on the electrocardiogram signal, sorts the music of each genre category and emotional subcategory according to the individual's heart rate and sympathetic nerve index feedback, classifies the music samples in the music sample library according to the sorting results, constructs a quantized variational autoencoder model based on the differently classified music samples to generate new music samples of similar genres, and generates and recommends suitable depression treatment music to users based on a recommendation confidence table.

[0005] With the rise of new technologies, a new evaluation system is also needed to evaluate the effectiveness of these technologies in treatment.

[0006] The paper "On the Use of AI for Generation of Functional Music to Improve Mental Health" explores the application of AIGC in music therapy. It combines GSR with a rating scale to construct an evaluation system for a music therapy model. It concludes that the AIGC model can reduce the impact of familiarity on therapeutic effectiveness by continuously generating new audio.

[0007] However, the existing evaluation system has the following shortcomings: First, the existing evaluation system is mainly based on self-assessment scales and therapists' experience judgments. It lacks a deep integration of objective physiological data and musical characteristics. The quantitative scoring is based on subjective feelings, and there is no strict uniformity in the standards.

[0008] Secondly, existing objective indicators based on physiological signals mostly focus on a single physiological indicator, ignoring the impact of music's own characteristics such as melodic complexity and rhythmic regularity on therapeutic efficacy.

[0009] In addition, traditional methods are mostly static assessments, which cannot track the interactive changes between the patient's psychological state and music characteristics during the treatment process in real time, making it difficult to accurately quantify the therapeutic effect, and therefore difficult to deploy in an automated music therapy system. Summary of the Invention

[0010] The present invention provides an evaluation method and device for an AIGC music healing system, which can improve the accuracy of evaluation and help improve the adaptability of the system and the level of personalization of treatment.

[0011] An evaluation method for an AIGC music healing system includes the following steps: (1) During the music healing process using the AIGC music healing system, the listener's physiological signals are collected in real time, and the synchronization rate between the music signal and the physiological signal is calculated. ; (2) Build a melody prediction model based on the LSTM network, input the historical note sequence and output the probability distribution of the next note, and further calculate the melody predictability score ; (3) Obtain subjective and objective indicators during the music healing process, standardize the subjective and objective indicators, and then calculate the current tense , and calculate the consistency of subjective and objective indicators at the same time ; (4) Based on synchronization rate , melody predictability score , present tense and consistency Construct a dynamic weighted total evaluation function and evaluate the AIGC music healing system based on the results of the total evaluation function.

[0012] In step (1), the listener's physiological signals including heart rate HR, skin conductance GSR and brain wave EEG signals are collected.

[0013] In step (1), calculate the synchronization rate between the music signal and the physiological signal , the formula is as follows: ; Where, represents the total length of the time series being evaluated, is a plural unit, is the time step The phase difference between the music signal sequence and the physiological signal sequence is obtained; the phase of the music signal sequence is extracted by filtering and Hilbert transforming the music signal; the phase of the physiological signal sequence is obtained by the same transformation.

[0014] In step (2), the melody predictability score is calculated , the formula is: ; Where, represents the total number of music time steps; Indicates the input history note sequence After that, the next note is output The probability distribution of .

[0015] In step (3), subjective indicators include scale scores, and objective indicators include heart rate variability (HRV), electroencephalogram (EEG, band signal) and galvanic skin response (GSR).

[0016] In step (3), the current tense is calculated after the subjective and objective indicators are standardized. , the formula is: ; Where, 、 、 and They are standardized scale scores, heart rate variability, electroencephalogram, and galvanic skin response; is the weight coefficient.

[0017] In step (3), the consistency between subjective and objective indicators is calculated. , the formula is: ; ; ; ; Where, Represents the weight coefficient, which is used to balance the absolute difference and Pearson correlation coefficient between subjective and objective indicators. represents the standard deviation of the objective indicator, represents the standard deviation of the subjective indicator, It represents the covariance between objective indicators and subjective indicators, and the three constitute the Pearson correlation coefficient.

[0018] In step (4), the formula of the total evaluation function is: ; ; Where, 、 、 、 is the weight of each indicator. In the process of music healing, 、 、 、 Dynamic changes.

[0019] An evaluation device for an AIGC music healing system includes a memory and one or more processors. The memory stores executable code. When the one or more processors execute the executable code, they are used to implement the evaluation method of the AIGC music healing system.

[0020] Compared with the prior art, the present invention has the following beneficial effects: The present invention first analyzes the synchronization between the temporal characteristics of music and the listener's physiological signals. It then applies melodic predictability to the evaluation of music therapy systems, using melodic predictability to represent the complexity of the musical structure. It then determines the current tense through subjective and objective indicators, and designs a subjective-objective consistency index to quantify the objective evaluation (physiological indicators) and subjective scoring (scale). Finally, a dynamically weighted overall evaluation function is proposed based on synchronization rate, melodic predictability, current tense, and subjective-objective consistency indicators. The present method is more suitable for automatic music generation systems, such as AIGC, and can incorporate the characteristics of the music itself into the evaluation of the therapeutic effect of music, thereby improving the accuracy of the evaluation. Furthermore, dynamic, real-time analysis helps improve the adaptability of the system and the personalization of treatment, and helps enhance the interpretability of musical responses and therapeutic effects. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 This is a flow chart of an evaluation method for an AIGC music healing system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0022] The present invention will be described in further detail below with reference to the accompanying drawings and examples. It should be noted that the following examples are intended to facilitate understanding of the present invention and do not have any limiting effect on the present invention.

[0023] In the music healing system based on AIGC, the evaluation of treatment effects can be divided into two parts: 1. Evaluation of the generation process is primarily based on the characteristics of the generated music. Quantitative analysis is performed on the quality and characteristics (rhythm, melody, harmony) and structural indicators (melodic predictability) of the AIGC-generated music to ensure that the generated music meets the treatment goals.

[0024] 2. Evaluation of the treatment process. This evaluation process involves the person's perception of music. The evaluation is mainly based on physiological responses and subjective reports collected during the treatment process.

[0025] like Figure 1 As shown, an evaluation method of the AIGC music healing system includes the following steps: Step S1, calculation of synchronization rate.

[0026] The analysis is conducted using the synchronization between the temporal characteristics of music and the listener's physiological signals.

[0027] The listener's heart rate (HR), skin conductance (GSR) and electroencephalogram (EEG) signals are collected in real time through wearable devices, and their periodic characteristics (such as the RR interval of HR and the gamma band energy of EEG) are extracted.

[0028] At the same time, the filtered music signal is subjected to Hilbert transform to extract the instantaneous phase and define the synchronization rate as: ; in, Time step The phase difference between the music sequence and the physiological signal sequence is calculated, and the phase locking value can well represent the degree of phase similarity between the two time series. ,The closer the value is to 1, it indicates that the physiological signal and the music beat have established a good synchronization, and a better sense of immersion can make the treatment effect more significant.

[0029] Step S2: melody predictability analysis.

[0030] This study uses melodic predictability in the evaluation of music therapy systems, using it to represent the complexity of a piece's structure. Lower complexity helps foster a sense of peace of mind, but the therapeutic effect is more likely to decrease with increasing familiarity.

[0031] Build a melody prediction model based on the LSTM network and input the historical note sequence , output the next note The probability distribution of , define the predictability score Pred as: ; Where, Represents the total number of music time steps. Step S3, current tense.

[0032] Not only does the present tense frequently appear in explanations of music therapy, but a good working state is also associated with the present tense and the flow experience.

[0033] The present tense can be determined by subjective and objective indicators: ; in, 、 、 and They are standardized scale scores, heart rate variability, electroencephalogram and skin electrodermal response. is the corresponding weight coefficient. In practical applications, For the EEG Band signal. Step S4, physiological-scale consistency.

[0034] Checklists, self-assessment scales, and medical observations have long served as primary evaluation tools, providing a wealth of criteria for traditional psychological interventions. However, due to their subjective nature, these criteria are not suitable for direct deployment in automated systems. Therefore, it is necessary to examine the consistency between subjective feedback and objective physiological responses.

[0035] To quantify the consistency between objective evaluation (physiological indicators) and subjective ratings (scales), the following composite indicators were designed: ; ; ; ; Where, Represents the weight coefficient, which is used to balance the absolute difference and Pearson correlation coefficient between subjective and objective indicators. represents the standard deviation of the objective indicator, represents the standard deviation of the subjective indicator, It represents the covariance between objective and subjective indicators.

[0036] Step S5: constructing the overall evaluation function.

[0037] Based on synchronization rate, melody predictability, current tense and subjective and objective consistency indicators, a dynamic weighted overall evaluation function is proposed: ; ; Where, 、 、 、 is the weight of each indicator. In the process of music healing, 、 、 、 Dynamic changes.

[0038] Dynamic weight adjustment strategy: The importance of each indicator changes as treatment progresses. The initialization and rate of change of weights are shown in Table 1. The overall goal of this strategy is to focus on enhancing immersion and direct emotional response in the early stages of treatment and on comprehensive evaluation of treatment effectiveness in the later stages.

[0039] Table 1

[0040] Based on the same inventive principle, an embodiment of the present invention also provides an evaluation device for an AIGC music healing system, comprising a memory and one or more processors, wherein the memory stores executable code, and when one or more processors execute the executable code, it is used to implement the evaluation method for the AIGC music healing system mentioned in the above embodiment.

[0041] The embodiments described above provide a detailed description of the technical solutions and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, supplements and equivalent substitutions made within the scope of the principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An evaluation method for an AIGC music healing system, characterized in that: The following steps are involved: (1) During the music healing process using the AIGC music healing system, the listener's physiological signals are collected in real time, and the synchronization rate between the music signal and the physiological signal is calculated. ; (2) Build a melody prediction model based on the LSTM network, input the historical note sequence and output the probability distribution of the next note, and further calculate the melody predictability score ; (3) Obtain subjective and objective indicators during the music healing process, standardize the subjective and objective indicators, and then calculate the current tense , and calculate the consistency of subjective and objective indicators at the same time ; (4) Based on synchronization rate , melody predictability score , present tense and consistency Construct a dynamic weighted total evaluation function and evaluate the AIGC music healing system based on the results of the total evaluation function.

2. The evaluation method of the AIGC music healing system according to claim 1, characterized in that: In step (1), the listener's physiological signals including heart rate HR, skin conductance GSR and brain wave EEG signals are collected.

3. The evaluation method of the AIGC music healing system according to claim 1, characterized in that: In step (1), calculate the synchronization rate between the music signal and the physiological signal , the formula is as follows: ； Where, represents the total length of the time series being evaluated, is a plural unit, is the time step The phase difference between the music signal sequence and the physiological signal sequence is obtained; the phases of the music signal sequence and the physiological signal sequence are extracted by filtering and Hilbert transforming the signals.

4. The evaluation method of the AIGC music healing system according to claim 1, characterized in that: In step (2), the melody predictability score is calculated , the formula is: ； Where, represents the total number of music time steps; Indicates the input history note sequence After that, the next note is output The probability distribution of .

5. The evaluation method of the AIGC music healing system according to claim 1, characterized in that: In step (3), subjective indicators include scale scores, and objective indicators include heart rate variability (HRV), electroencephalogram (EEG), and galvanic skin response (GSR).

6. The evaluation method of the AIGC music healing system according to claim 5, characterized in that: In step (3), the current tense is calculated after the subjective and objective indicators are standardized. , the formula is: ； Where, 、 、 and They are standardized scale scores, heart rate variability, electroencephalogram, and galvanic skin response; is the weight coefficient.

7. The evaluation method of the AIGC music healing system according to claim 6, characterized in that: In step (3), the consistency between subjective and objective indicators is calculated. , the formula is: ； ； ； ； Where, Represents the weight coefficient, which is used to balance the absolute difference between subjective and objective indicators and the proportion of the Pearson correlation coefficient. represents the standard deviation of the objective indicator, represents the standard deviation of the subjective indicator, It represents the covariance between objective indicators and subjective indicators, and the three constitute the Pearson correlation coefficient.

8. The evaluation method of the AIGC music healing system according to claim 6, characterized in that: In step (4), the formula of the total evaluation function is: ； ； Where, 、 、 、 is the weight of each indicator. In the process of music healing, 、 、 、 Dynamic changes.

9. An evaluation device for an AIGC music healing system, characterized in that: It includes a memory and one or more processors, wherein the memory stores executable code, and when the one or more processors execute the executable code, it is used to implement the evaluation method of the AIGC music healing system according to any one of claims 1 to 8.

Images