Student psychological health dynamic assessment method and system based on deep learning

Abstract

The invention relates to the technical field of data processing, and particularly discloses a student psychological health dynamic assessment method and system based on deep learning, and the method comprises the steps: collecting student psychological health data through a target application platform, monitoring the data processing process in real time, and analyzing video transmission quality parameters through deep learning. A video transmission process is dynamically adjusted to ensure accurate transmission of data; meanwhile, audio synchronization parameters are collected and evaluated, the audio synchronization process is adjusted, video and audio fluctuation parameters are processed, and video and audio stability is optimized; and finally, automatically extracting adjusted data feature information, and outputting student psychological health information in combination with a deep learning model. According to the method, multi-parameter analysis and deep learning technologies are fused, the accuracy and real-time performance of evaluation are effectively improved, and a scientific and efficient solution is provided for student mental health monitoring and intervention.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, in particular to a student mental health dynamic evaluation method and system based on deep learning. BACKGROUND

[0002] Students are in a critical period of physical and mental development, and mental health is crucial to their growth, learning and future development. Deep learning models adapt to different groups through transfer learning, can handle large amounts of multi-source data, reduce dependence on large-scale labeled data, and through automatic feature extraction and analysis of these data, deep learning can mine potential information hidden in the data, more comprehensively understand the mental health status of students, and adapt to different types of data and complex psychological characteristics, improving the accuracy of evaluation.

[0003] For example, the invention patent application with publication number CN105740224A discloses a user psychological warning method and device based on text analysis, which includes obtaining all speech text content related to the user; then calculating the user's emotional index by using a sentence-level emotional analysis method on the text; at the same time, the user's text content is classified by using a method based on text cosine similarity; then the user's emotional index and emotional type are used for psychological evaluation and warning; finally, the user's psychological evaluation and warning results are displayed.

[0004] For example, the invention patent application with publication number CN118965176A discloses a user emotion recognition and psychological analysis method based on a semantic mental health model, which includes: obtaining the type of the current logged-in student's interactive scene; determining the scene interaction event according to the type of the interactive scene; triggering student interaction data according to the scene interaction event; performing semantic analysis on the student interaction data through a pre-trained semantic mental health model, and calculating the probability of the psychological characteristics in each psychological characteristic type; determining the next round of triggered scene interaction event according to the student psychological analysis report.

[0005] However, in the process of implementing the embodiments of the present application, it is found that the above-mentioned technology at least has the following technical problems: the existing student mental health evaluation method has uneven data quality and lacks deep processing. On the one hand, the voice collection environment is complex, and the background noise and equipment differences in different scenes such as classrooms and consultation rooms lead to uneven voice signal quality, affecting the accuracy of feature extraction. On the other hand, voice data labeling relies on manual work and the standard is not unified, which limits the quality of training data, and further leads to insufficient accuracy of mental health evaluation. SUMMARY

[0006] In view of the deficiencies of the prior art, the present application provides a student mental health dynamic evaluation method and system based on deep learning, which can effectively solve the problems involved in the background technology.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: The first aspect of the present invention provides a dynamic assessment method for student mental health based on deep learning, including: step one, collecting student mental health data through a target application platform, marking it as target data, monitoring the processing process of the target data, obtaining and analyzing video transmission quality parameters, and dynamically judging whether to adjust the video transmission process of the target data based on deep learning; step two, collecting and evaluating audio synchronization parameters, and judging whether to adjust the audio synchronization process of the target data; step three, obtaining and processing audio and video fluctuation parameters, and judging whether to adjust the audio and video fluctuation stability of the target data; step four, automatically extracting the characteristic information of the adjusted target data, thereby obtaining the characteristic information of the student mental health data.

[0008] As a further method, it is determined whether to adjust the video transmission process of the target data. The specific judgment process is: comparing the video data transmission quality coefficient with the video data transmission quality threshold; if the video data transmission quality coefficient is greater than or equal to the video data transmission quality threshold, it is determined that the video transmission process of the target data is not adjusted; if the video data transmission quality coefficient is less than the video data transmission quality threshold, it is determined that the video transmission process of the target data is adjusted.

[0009] As a further method, the video transmission process of the target data is adjusted. The specific adjustment process is: based on the video data transmission quality coefficient and the video data transmission quality threshold, the video data transmission quality coefficient deviation value is obtained, and the increase coefficient is matched based on the video data transmission quality coefficient deviation value, so as to increase the buffer area; and the video data transmission quality coefficient improvement rate within the adjustment period is monitored, and compared with the video data transmission quality coefficient improvement rate limit value to obtain the comparison result, and the duration corresponding to the adjustment period is adjusted according to the comparison result.

[0010] As a further method, it is determined whether to adjust the audio synchronization process of the target data. The specific judgment process is: compare the audio synchronization coefficient with the audio synchronization system threshold. If the audio synchronization coefficient is greater than or equal to the audio synchronization system threshold, it is determined that the audio synchronization process of the target data is not adjusted; if the audio synchronization coefficient is less than the audio synchronization system threshold, it is determined that the audio synchronization process of the target data is adjusted.

[0011] As a further method, the audio synchronization process of the target data is adjusted, and the adjustment process is as follows: based on the audio synchronization threshold and the audio synchronization coefficient, an audio synchronization coefficient deviation value is obtained, and compared with an audio synchronization coefficient deviation threshold; if the audio synchronization coefficient deviation value is greater than or equal to the audio synchronization coefficient deviation threshold, the audio sampling rate is increased based on the audio synchronization coefficient deviation value, a secondary audio synchronization coefficient is obtained, and compared with the audio synchronization threshold, if the secondary audio synchronization coefficient is still less than the audio synchronization threshold, secondary adjustment is performed; if the secondary audio synchronization coefficient is greater than or equal to the audio synchronization threshold, no secondary adjustment is performed; if the audio synchronization coefficient deviation value is less than the audio synchronization coefficient deviation threshold, the audio frame interval duration is reduced based on the audio synchronization coefficient deviation value.

[0012] As a further method, it is judged whether to adjust the audio-video fluctuation stability of the target data, and the judgment process is as follows: the audio-video fluctuation index is compared with an audio-video fluctuation index threshold; if the audio-video fluctuation index is less than or equal to the audio-video fluctuation index threshold, it is judged that the initial running parameter is not adjusted, and the audio-video fluctuation stability of the target data is not adjusted; if the audio-video fluctuation index is greater than the audio-video fluctuation index threshold, it is judged that the audio-video fluctuation stability of the target data is adjusted, and the adjustment process is as follows: based on the audio-video fluctuation index and the audio-video fluctuation index threshold, an audio-video fluctuation index deviation value is obtained, and the encoding rate is adjusted based on the audio-video fluctuation index deviation value.

[0013] The second aspect of the present application provides a student mental health dynamic evaluation system based on deep learning, comprising: a video transmission quality adjustment module for collecting student mental health data through a target application platform and marking as target data, monitoring the processing process of the target data, obtaining and analyzing video transmission quality parameters, and dynamically judging whether to adjust the video transmission process of the target data based on deep learning; an audio synchronization adjustment module for collecting and evaluating audio synchronization parameters, and judging whether to adjust the audio synchronization process of the target data; an audio-video fluctuation adjustment module for obtaining and processing audio-video fluctuation parameters, and judging whether to adjust the audio-video fluctuation stability of the target data; a feature extraction module for automatically extracting feature information of the adjusted target data, so as to obtain feature information of the student mental health data.

[0014] Compared with the prior art, the embodiments of the present application have at least the following advantages or beneficial effects:

[0015] (1) The application provides a student mental health dynamic evaluation method and system based on deep learning, which collects comprehensive mental health data on the target application platform, monitors the processing in real time, dynamically analyzes the video transmission quality parameters with the help of deep learning, ensures the complete and accurate transmission of video information, accurately collects and evaluates audio synchronization parameters, realizes the accurate matching of audio and video information, deeply processes the audio and video fluctuation parameters, reduces data interference and improves data quality, automatically extracts the data characteristics after adjustment, and outputs high-reliability student mental health information. The whole process realizes automation and intelligent operation, effectively eliminates interference factors, accurately captures key information, provides reliable basis for student mental health evaluation, and significantly improves the accuracy of evaluation results.

[0016] (2) The target application platform collects student mental health data and marks it as target data, monitors the processing, obtains and analyzes video transmission quality parameters, and dynamically judges whether to adjust the video transmission process based on deep learning. This can ensure that the collected data is comprehensive and true, and in the transmission link, with the powerful analysis capability of deep learning, problems that may affect the integrity and accuracy of data in video transmission can be found in real time and accurately, and transmission strategies can be adjusted in time to avoid distortion of mental health data due to poor video transmission quality, laying a reliable data foundation for subsequent analysis.

[0017] (3) By collecting and evaluating audio synchronization parameters, it is determined whether to adjust the audio synchronization process of the target data. Since audio information plays an important role in student mental health evaluation, such as voice tone, environmental sound effect, etc., which can reflect the psychological state of students, by accurately collecting and evaluating audio synchronization parameters, the accurate synchronization of audio and video can be ensured, and the situation of sound and picture mismatch can be avoided, ensuring that the sound, environment sound, etc. Audio information emitted by students in various scenes can be accurately and timely recorded and analyzed, making the evaluation results more comprehensive and accurate.

[0018] (4) In the process of data transmission and processing, the video and audio signals are easily disturbed by various factors and fluctuate, which may affect the accurate judgment of student mental health. By obtaining and processing the fluctuation parameters, the video picture is more stable, the audio sound is clearer, the misjudgment caused by signal fluctuation is reduced, the quality and reliability of the data are further improved, and the accuracy of student mental health evaluation is improved. BRIEF DESCRIPTION OF DRAWINGS

[0019] The application is further illustrated by the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the application. For ordinary skilled in the art, other drawings can be obtained without creative labor on the basis of the following drawings.

[0020] Figure 1A flowchart of the method steps of the present application.

[0021] Figure 2 A first detailed flowchart of the method steps of the present application.

[0022] Figure 3 A second detailed flowchart of the method steps of the present application.

[0023] Figure 4 A third detailed flowchart of the method steps of the present application.

[0024] Figure 5 A schematic diagram of the system module connection of the present application. DETAILED DESCRIPTION

[0025] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0026] Referring to Figure 1 The first aspect of the present application provides a deep learning-based dynamic evaluation method for student mental health, which comprises the following steps: step one, collecting student mental health data through a target application platform and marking it as target data, monitoring the processing process of the target data, obtaining and analyzing video transmission quality parameters, and dynamically determining whether to adjust the video transmission process of the target data based on deep learning.

[0027] Specifically, the video transmission quality parameters are obtained and analyzed, and the specific analysis process is as follows: the video transmission quality parameters include the peak-to-average power ratio of the video transmission channel, the channel fading depth factor of the video transmission channel, and the inter-symbol interference power factor of the video transmission channel.

[0028] It should be explained that the peak-to-average power ratio of the video transmission channel represents the proportional relationship between the peak power of the video transmission channel and the average power of the video transmission channel; the channel fading depth factor of the video transmission channel represents the proportional relationship between the channel fading depth of the video transmission channel and the channel fading depth limit value; and the inter-symbol interference power factor of the video transmission channel represents the proportional relationship between the inter-symbol interference power of the video transmission channel and the inter-symbol interference power limit value.

[0029] By introducing influence coefficients, we can quantify the influence of the peak-to-average power ratio of the video transmission channel on the video data transmission quality coefficient, the influence of the channel fading depth factor of the video transmission channel on the video data transmission quality coefficient, and the influence of the inter-symbol interference power factor of the video transmission channel on the video data transmission quality coefficient. By coupling each influence degree, we can obtain the video data transmission quality coefficient. The video data transmission quality coefficient represents the quality of the video data during the transmission process. The specific evaluation method is:

[0030] ;

[0031] ;

[0032] ;

[0033] ;

[0034] Where, is the video data transmission quality coefficient, is the peak-to-average power ratio of the video transmission channel, is the peak power of the video transmission channel, is the average power of the video transmission channel, is the channel fading depth factor of the video transmission channel, is the channel fading depth of the video transmission channel, is the channel fading depth limit value preset in the video database, is the inter-symbol interference power factor of the video transmission channel, is the inter-symbol interference power of the video transmission channel, is the inter-symbol interference power threshold value preset in the video database, is the influence coefficient corresponding to the peak-to-average power ratio preset in the video database, is the influence coefficient corresponding to the channel fading depth factor preset in the video database, It is the influence coefficient corresponding to the inter-symbol interference power factor preset in the video database.

[0035] It needs to be explained that the peak power of the above video transmission channel represents the maximum instantaneous power value that the transmission channel can reach during video transmission, which can be measured using a power meter; the average power of the above video transmission channel represents the average value of the transmission power on the video transmission channel, which can also be measured using a power meter; the channel fading depth of the above video transmission channel represents the degree of reduction of signal strength due to channel fading, which can be obtained by using a pilot signal or a channel estimation model based on deep learning; the inter-symbol interference power of the above video transmission channel represents the energy size of the physical quantity that the current code element signal waveform is extended to the adjacent code element time interval to interfere with its judgment, which can be calculated according to the model and signal transmission characteristics of the communication system through mathematical derivation (such as convolution operation).

[0036] The channel fading depth limit value represents the maximum value of the allowed channel fading depth in the video database; the inter-symbol interference power limit value represents the maximum value of the allowed inter-symbol interference power in the video database.

[0037] The influence coefficient corresponding to the peak-to-average power ratio represents the influence degree of a unit value of the peak-to-average power ratio on the video data transmission quality coefficient; the influence coefficient corresponding to the channel fading depth factor represents the influence degree of a unit value change of the channel fading depth factor on the video data transmission quality coefficient; the influence coefficient corresponding to the inter-symbol interference power factor represents the influence degree of a unit value change of the inter-symbol interference power factor on the video data transmission quality coefficient; the video database stores the corresponding relationship between the peak-to-average power ratio and the influence coefficient corresponding to the peak-to-average power ratio, the corresponding relationship between the channel fading depth factor and the influence coefficient corresponding to the channel fading depth factor, and the corresponding relationship between the inter-symbol interference power factor and the influence coefficient corresponding to the inter-symbol interference power factor; for example, input the peak-to-average power ratio, the channel fading depth factor, and the inter-symbol interference power factor into the video database, and the video database can match the influence coefficient corresponding to the peak-to-average power ratio, the influence coefficient corresponding to the channel fading depth factor, and the influence coefficient corresponding to the inter-symbol interference power factor, with a value range of 0 to 1.

[0038] A higher peak-to-average power ratio means that the signal power fluctuates greatly, which deepens the influence of channel fading on the signal and makes the signal quality received by the receiving end worse; when the signal peak power is too high, the sampling point may be misjudged as the signal of the adjacent code element, resulting in an increase in inter-symbol interference power and affecting the correct demodulation of the signal; the increase in channel fading depth will cause changes in the amplitude and phase of the received signal, resulting in distortion of the signal waveform and increasing the degree of inter-symbol interference, which leads to an increase in inter-symbol interference power.

[0039] The peak-to-average power ratio is high, which means that the signal power fluctuates greatly, which easily leads to loss of details of the video signal, mosaic or blurred picture and other problems, and reduces the video data transmission quality; the greater the channel fading depth, the more serious the signal strength attenuation, which causes excessive attenuation of some frequency components, resulting in frequency distortion of the video signal, affecting the clarity and color restoration of the picture, and thus reducing the video data transmission quality; the greater the inter-symbol interference power, the more serious the interference between adjacent code elements. This will cause errors in the decision of the code elements at the receiving end, increase the bit error rate, and thus reduce the video data transmission quality.

[0040] Specifically, whether to adjust the video transmission process of the target data is determined by comparing the video data transmission quality coefficient with a video data transmission quality threshold value; if the video data transmission quality coefficient is greater than or equal to the video data transmission quality threshold value, it is determined that the video transmission process of the target data is not adjusted; if the video data transmission quality coefficient is less than the video data transmission quality threshold value, it is determined that the video transmission process of the target data is adjusted.

[0041] It should be explained that the above-mentioned video data transmission quality threshold value represents the minimum value allowed for the video data transmission quality extracted from the video database.

[0042] Further, the video transmission process of the target data is adjusted, and the specific adjustment process is as follows: based on the video data transmission quality coefficient and the video data transmission quality threshold value, a video data transmission quality coefficient deviation value is obtained, a magnification coefficient is matched based on the video data transmission quality coefficient deviation value, so as to increase the size of the buffer area; and the video data transmission quality coefficient improvement rate in the adjustment period is monitored and compared with the video data transmission quality coefficient improvement rate limit value to obtain a comparison result, and the time length corresponding to the adjustment period is adjusted according to the comparison result.

[0043] It should be explained that the above-mentioned video data transmission quality coefficient deviation value refers to the difference between the video data transmission quality coefficient and the video data transmission quality coefficient threshold value, and the processing result is the video data transmission quality coefficient deviation value; the above-mentioned increasing adjustment of the buffer area is specifically as follows: the buffer area magnification coefficient corresponding to each video data transmission quality coefficient deviation value interval in the video database is stored, the obtained video data transmission quality coefficient deviation value is input into the video database, the video database can match the corresponding video data transmission quality coefficient deviation value interval, and then the magnification coefficient corresponding to the interval is the buffer area magnification coefficient; the obtained buffer area magnification coefficient is multiplied by the size of the original buffer area, and the result obtained is the size of the buffer area to be adjusted; by increasing the size of the buffer area, data transmission jitter can be alleviated and the influence of packet loss can be reduced; the above-mentioned buffer area magnification coefficient is greater than 1, indicating that the buffer area needs to be increased by a multiple value.

[0044] The video data transmission quality coefficient improvement rate within the aforementioned period refers to the video data transmission quality coefficient improvement rate within the period, which is obtained by subtracting the video data transmission quality coefficient at the beginning of the period from the video data transmission quality coefficient at the end of the period. The result is divided by the period duration to obtain the video data transmission quality coefficient improvement rate within the period, indicating the degree of improvement in video data transmission quality after increasing the buffer. The aforementioned video data transmission quality coefficient improvement rate threshold value indicates the maximum value of the video data transmission quality coefficient improvement rate extracted from the video database. The aforementioned adjustment of the period duration corresponding to the comparison result is specifically performed as follows: if the video data transmission quality coefficient improvement rate is greater than or equal to the video data transmission quality coefficient improvement rate threshold value, it indicates that the video data transmission quality has improved well, and the adjustment period duration needs to be shortened to avoid increased delay caused by an excessively large buffer interval. If the video data transmission quality coefficient improvement rate is less than the video data transmission quality coefficient improvement rate threshold value, it indicates that the video data transmission quality has improved poorly, and the adjustment period duration needs to be extended to avoid increased freeze rate caused by an excessively small buffer interval. The aforementioned adjustment period duration is specifically adjusted as follows: the video database stores the period duration corresponding to each video data transmission quality coefficient improvement rate. The obtained video data transmission quality coefficient improvement rate stored in the video database is input into the video database to match the corresponding adjustment period duration.

[0045] In a specific embodiment, students' mental health data is collected through a target application platform and marked as target data. At the same time, its processing process is monitored, video transmission quality parameters are obtained and analyzed, and deep learning is used to dynamically determine whether to adjust the video transmission process. This ensures that the collected data is comprehensive and true, and in the transmission link, with the help of the powerful analytical capabilities of deep learning, problems in video transmission that may affect data integrity and accuracy can be discovered in real time and accurately, and transmission strategies can be adjusted in a timely manner to avoid distortion of mental health data due to poor video transmission quality, thereby laying a reliable data foundation for subsequent analysis.

[0046] Step 2: Collect and evaluate audio synchronization parameters, and determine whether to adjust the audio synchronization process of the target data.

[0047] Specifically, audio synchronization parameters are collected and evaluated. The specific evaluation process is as follows: the audio synchronization parameters include the zero-crossing rate difference of the audio signal, the information entropy difference of the audio signal, and the spectral kurtosis difference of the audio signal.

[0048] It should be explained that the zero-crossing rate difference of the above-mentioned audio signal represents the absolute value of the difference between the audio zero-crossing rate and the video zero-crossing rate; the information entropy difference of the above-mentioned audio signal represents the absolute value of the difference between the audio information entropy and the video information entropy; the spectral kurtosis difference of the above-mentioned audio signal represents the absolute value of the difference between the audio spectral kurtosis and the video spectral kurtosis; the above-mentioned audio zero-crossing rate, video zero-crossing rate, audio information entropy, video information entropy, audio spectral kurtosis and video spectral kurtosis need to be normalized using the Z-score normalization method.

[0049] By introducing influence coefficients, we quantify the influence of the zero-crossing rate difference of the audio signal on the audio synchronization coefficient, the influence of the information entropy difference of the audio signal on the audio synchronization coefficient, the influence of the spectral kurtosis difference of the audio signal on the audio synchronization coefficient, and the influence of the video data transmission quality coefficient on the audio synchronization coefficient. By coupling each influence degree, we can obtain the audio synchronization coefficient, where the audio synchronization coefficient represents the degree of synchronization between audio and video. The specific evaluation method is:

[0050] ;

[0051] ;

[0052] ;

[0053] ;

[0054] Where, is the audio synchronization coefficient, is the video data transmission quality coefficient, is the zero-crossing rate difference of the audio signal, is the audio zero-crossing rate, is the video zero-crossing rate, is the information entropy difference of the audio signal, is the audio information entropy, is the video information entropy, is the spectral kurtosis difference of the audio signal, is the audio spectrum kurtosis, is the video spectrum kurtosis, is the influence coefficient corresponding to the zero-crossing rate difference preset in the video database, is the influence coefficient corresponding to the information entropy difference preset in the video database, is the influence coefficient corresponding to the spectral kurtosis difference value preset in the video database, is the influence coefficient corresponding to the video transmission quality coefficient preset in the video database, 、 、 are all constants.

[0055] It should be explained that the above-mentioned audio zero-crossing rate represents the number of times the audio signal crosses the zero level per unit time. By sampling the audio signal, a discrete audio sample sequence is obtained, and then the sample sequence is traversed to count the number of times the signs of adjacent sample values ​​change (i.e., from positive to negative or from negative to positive). The number of times the sign of the adjacent sample values ​​changes is counted, and then divided by the number of sampling points and the time interval, the audio zero-crossing rate can be obtained. The above-mentioned video zero-crossing rate represents the frequency of signal changes in a certain dimension of the video image (such as brightness value). This can be achieved by converting the video frame into a grayscale image (if it is a color video), and then traversing the pixel values ​​by rows or columns to count the adjacent pixel values. The number of times the symbol changes is normalized in combination with the size and time information of the image to obtain the video zero-crossing rate; the above-mentioned audio information entropy represents the amount of information or uncertainty contained in the audio signal, which can be obtained through the information entropy calculation formula; the above-mentioned video information entropy represents the richness and uncertainty of information in the video image, which can be obtained through the information entropy calculation formula; the above-mentioned audio spectral kurtosis represents the flatness or sharpness of the audio signal spectrum, which can be obtained through the wavelet transform method; the above-mentioned video spectral kurtosis represents the distribution of video image spectrum energy, which is obtained through the local binary pattern and spectrum analysis method.

[0056] The influence coefficient corresponding to the above zero-crossing rate difference indicates the degree of influence of the unit value of the zero-crossing rate difference on the audio synchronization coefficient; the influence coefficient corresponding to the above information entropy difference indicates the degree of influence of the unit value of the information entropy difference on the audio synchronization coefficient; the influence coefficient corresponding to the above spectral kurtosis difference indicates the degree of influence of the unit value of the spectral kurtosis difference on the audio synchronization coefficient; the influence coefficient corresponding to the above video transmission quality coefficient indicates the degree of influence of the unit value of the video transmission quality coefficient on the audio synchronization coefficient; the video database stores the correspondence between the zero-crossing rate difference and its corresponding influence coefficient, the information entropy difference and its corresponding influence coefficient. The correspondence between the influence coefficients corresponding to the information entropy difference, the correspondence between the spectral kurtosis difference and the influence coefficients corresponding to its corresponding spectral kurtosis difference, and the correspondence between the influence coefficients corresponding to the video transmission quality coefficient and the corresponding video transmission quality coefficient; for example, the zero-crossing rate difference, the information entropy difference, the spectral kurtosis difference and the video transmission quality coefficient are input into the video database, and the video database can match the influence coefficients corresponding to the zero-crossing rate difference, the influence coefficients corresponding to the information entropy difference, the influence coefficients corresponding to the spectral kurtosis difference and the influence coefficients corresponding to the video transmission quality coefficient, and the value range is between 0 and 1.

[0057] A large difference in zero-crossing rate indicates that the audio and video have a large difference in the frequency of signal changes. A large difference in information entropy means that the audio and video have a large difference in the richness and clutter of information. The information entropy of audio with frequent signal changes will also be relatively high, that is, when the zero-crossing rate difference is large, the information entropy difference is also large; when the zero-crossing rate difference is large, it implies that there are differences in their frequency components and energy distribution, which in turn leads to a large difference in spectral kurtosis; when the information entropy difference is large, it means that the audio and video have a large difference in information complexity and uniformity of energy distribution, which will lead to a large difference in spectral kurtosis.

[0058] The smaller the difference in zero-crossing rate, the more similar the changing rhythm of the audio and video signals on the time axis, that is, the signal changes of audio and video are more synchronized in time, and the audio synchronization coefficient may be higher; the smaller the difference in audio information entropy, the more similar the amount of information and complexity contained in the audio and video are, which implies that the audio and video are more matched in content and rhythm, and thus the audio synchronization coefficient is higher; the smaller the difference in audio spectral kurtosis, the more similar the audio and video are in the sharpness of the spectrum and the concentration of energy distribution, which reflects that the audio and video are more synchronized in frequency characteristics and rhythm, and the audio synchronization coefficient may be higher; a high video transmission quality coefficient means that the video signal is well maintained during the transmission process, with low distortion and less packet loss, which provides a good foundation for the synchronization of audio and video signals.

[0059] Furthermore, it is determined whether to adjust the audio synchronization process of the target data. The specific judgment process is: compare the audio synchronization coefficient with the audio synchronization coefficient threshold. If the audio synchronization coefficient is greater than or equal to the audio synchronization coefficient threshold, it is determined that the audio synchronization process of the target data is not adjusted; if the audio synchronization coefficient is less than the audio synchronization coefficient threshold, it is determined that the audio synchronization process of the target data is adjusted.

[0060] It should be explained that the above audio synchronization coefficient threshold represents the minimum value of the audio synchronization coefficient extracted from the video database.

[0061] Specifically, the audio synchronization process of the target data is adjusted. The specific adjustment process is as follows: based on the audio synchronization threshold and the audio synchronization coefficient, an audio synchronization coefficient deviation value is obtained, and the audio synchronization coefficient deviation threshold is compared; the above-mentioned audio synchronization coefficient deviation value refers to the result of processing the difference between the audio synchronization threshold and the audio synchronization coefficient; if the audio synchronization coefficient deviation value is greater than or equal to the audio synchronization coefficient deviation threshold, the audio sampling rate is increased based on the audio synchronization coefficient deviation value, a secondary audio synchronization coefficient is obtained, and the audio synchronization coefficient is compared with the audio synchronization threshold; if the secondary audio synchronization coefficient is still less than the audio synchronization threshold, a secondary adjustment is performed; if the secondary audio synchronization coefficient is greater than or equal to the audio synchronization threshold, no secondary adjustment is performed;

[0062] If the audio synchronization coefficient deviation value is less than the audio synchronization coefficient deviation threshold, the audio frame interval length is reduced based on the audio synchronization coefficient deviation value, so that the audio data can be sent more compactly, reducing the waiting time of the receiving end, thereby improving the audio synchronization coefficient.

[0063] It should be explained that the above-mentioned audio synchronization coefficient deviation threshold represents the maximum value of the audio synchronization coefficient deviation allowed in the video database; the above-mentioned process of increasing the audio sampling rate based on the audio synchronization coefficient deviation value is as follows: the sampling frequency increase coefficient corresponding to each audio synchronization coefficient deviation value interval is stored in the video database, and the obtained audio synchronization coefficient deviation value is input into the video database. The video database can match the corresponding audio synchronization coefficient deviation value interval, and the sampling frequency increase coefficient corresponding to the interval is the increase coefficient that needs to be adjusted. The sampling frequency increase coefficient is multiplied by the original sampling frequency, and the result is the sampling frequency that needs to be increased. The above-mentioned sampling frequency increase coefficient is greater than 1, which indicates that the audio sampling rate needs to be increased by a multiple; the above-mentioned Secondary adjustment refers to further switching to an efficient encoding algorithm (such as converting from PCM encoding to Opus encoding). The above-mentioned reduction of the audio frame interval duration based on the audio synchronization coefficient deviation value is specifically carried out as follows: the audio frame interval duration reduction coefficient corresponding to each audio synchronization coefficient deviation value interval is stored in the video database, and the obtained audio synchronization coefficient deviation value is input into the video database. The video database can then match the corresponding audio synchronization coefficient deviation value interval. The audio frame interval duration reduction coefficient corresponding to the interval is the coefficient that needs to be adjusted. The coefficient is multiplied by the original audio frame interval duration, and the result is the audio frame interval duration that needs to be adjusted. The above-mentioned audio frame interval duration reduction coefficient is less than 1, which indicates the proportion by which the audio frame interval duration needs to be reduced.

[0064] In a specific embodiment, by collecting and evaluating audio synchronization parameters, it is determined whether to adjust the audio synchronization process of the target data; since audio information plays an important role in the mental health assessment of students, such as voice intonation, environmental sound effects, etc., which can reflect the psychological state of students, by accurately collecting and evaluating audio synchronization parameters, accurate synchronization of audio and video can be guaranteed, avoiding the mismatch between sound and picture, and ensuring that the audio information such as the sounds emitted by students in various scenarios and the sound effects of the environment in which they are located can be accurately and timely recorded and analyzed, making the evaluation results more comprehensive and accurate.

[0065] Step 3: Obtain and process audio-visual fluctuation parameters to determine whether to adjust the audio-visual fluctuation stability of the target data.

[0066] Specifically, the audio-visual fluctuation parameters are obtained and processed, and the specific processing process is as follows: the audio-visual fluctuation parameters include the target data network packet disorder rate factor and the target data fractal dimension factor.

[0067] It should be explained that the above-mentioned target data network packet disorder rate factor represents the proportional relationship between the network packet disorder rate fluctuation value of the target data and the network packet disorder rate fluctuation limit value; the above-mentioned target data fractal dimension factor represents the absolute value of the difference between the audio fractal dimension and the video fractal dimension; the above-mentioned audio fractal dimension value and the video fractal dimension value need to be normalized using the Z-score normalization method.

[0068] By introducing influence coefficients, we can quantify the influence of the target data network packet disorder rate factor on the audio-visual fluctuation index, the influence of the target data fractal dimension factor on the audio-visual fluctuation index, the influence of the video data transmission quality coefficient fluctuation value within a period on the audio-visual fluctuation index, and the influence of the audio synchronization coefficient fluctuation value within a period on the audio-visual fluctuation index. By coupling each influence degree, we can obtain the audio-visual fluctuation index. The audio-visual fluctuation index represents the fluctuation of the synchronization between video and audio. The specific evaluation method is as follows:

[0069] ;

[0070] ;

[0071] ;

[0072] Where, is the audiovisual fluctuation index, is the fluctuation value of the video data transmission quality coefficient within the period, is the fluctuation value of the audio synchronization coefficient within the period, is the target data network packet disorder rate factor, is the disorder rate of the network packets of the target data, The network packet disorder rate threshold value preset in the video database, is the fractal dimension factor within the target data period, is the audio fractal dimension, is the video fractal dimension, is the influence coefficient corresponding to the network packet disorder rate factor preset in the video database, is the influence coefficient corresponding to the fractal dimension factor preset in the video database, The influence coefficient corresponding to the fluctuation value of the video data transmission quality coefficient within a preset period in the video database, is the influence coefficient corresponding to the fluctuation value of the audio synchronization coefficient within the preset period in the video database, is a constant.

[0073] It should be explained that the fluctuation value of the video data transmission quality coefficient within the above-mentioned period refers to the absolute value of the difference between the video data transmission quality coefficient at the beginning of the period and the video data transmission quality coefficient at the end of the period; the fluctuation value of the audio synchronization coefficient within the above-mentioned period refers to the absolute value of the difference between the audio synchronization coefficient at the beginning of the period and the audio synchronization coefficient at the end of the period; the above-mentioned network packet disorder rate indicates the proportion of data packets that do not arrive at the receiving end in the order of sending during the network data transmission process, which can usually be obtained through network monitoring tools; the above-mentioned audio fractal dimension indicates the complexity and irregularity of the audio signal, which can be obtained through the box dimension method; the above-mentioned video fractal dimension indicates the complexity and self-similarity characteristics of the video image sequence, which can be obtained through the differential box dimension method; the above-mentioned network packet disorder rate limit value indicates the maximum value of the network packet disorder rate allowed in the video database.

[0074] The influence coefficient corresponding to the above-mentioned network packet disorder rate factor indicates the degree of influence of the unit value of the network packet disorder rate factor on the audio-visual fluctuation index; the influence coefficient corresponding to the above-mentioned fractal dimension factor indicates the degree of influence of the unit value of the fractal dimension factor on the audio-visual fluctuation index; the influence coefficient corresponding to the fluctuation value of the video data transmission quality coefficient within the above-mentioned period indicates the degree of influence of the unit value of the fluctuation value of the video data transmission quality coefficient within the period on the audio-visual fluctuation index; the influence coefficient corresponding to the fluctuation value of the audio synchronization coefficient within the above-mentioned period indicates the degree of influence of the unit value of the fluctuation value of the audio synchronization coefficient within the period on the audio-visual fluctuation index; the device database stores the correspondence between the network packet disorder rate factor and its corresponding influence coefficient, and the fractal dimension factor and its corresponding fractal dimension factor. The corresponding relationship between the influence coefficients, the corresponding relationship between the fluctuation value of the video data transmission quality coefficient within the period and the influence coefficient corresponding to the fluctuation value of the video data transmission quality coefficient within the period, and the corresponding relationship between the fluctuation value of the audio synchronization coefficient within the period and the influence coefficient corresponding to the fluctuation value of the audio synchronization coefficient within the period; for example, the network packet disorder rate factor, the fractal dimension factor, the fluctuation value of the video data transmission quality coefficient within the period, and the fluctuation value of the video data transmission quality coefficient within the period are input into the video database, and the video database can match the influence coefficient corresponding to the network packet disorder rate factor, the influence coefficient corresponding to the fractal dimension factor, the fluctuation value of the video data transmission quality coefficient within the period, and the influence coefficient corresponding to the fluctuation value of the audio synchronization coefficient within the period, and the value range is between 0 and 1.

[0075] The network packet disorder rate is too high, which can cause the transmission order of video and audio data to be chaotic, and thus increase the video data transmission quality coefficient fluctuation value. In addition, because the quality of audio and video changes, the fractal dimension factor is also affected. At the same time, the audio and video are out of synchronization, which increases the audio synchronization coefficient fluctuation value. The increase of the video data transmission quality coefficient fluctuation value indicates that there is a problem in the video transmission process, which reflects that the network packet disorder rate is too high. At the same time, the audio synchronization is also out of order, which increases the audio synchronization coefficient fluctuation value. The change of the video quality also affects the video fractal dimension value, and thus affects the fractal dimension factor.

[0076] The higher the network packet disorder rate, the worse the stability of network transmission, and the greater the possibility of video and audio fluctuation. Therefore, the lower the network packet disorder rate factor, the lower the audio and video fluctuation index. The lower the audio fractal dimension, the less likely it is to be affected by various factors during transmission and processing, and thus the lower the audio and video fluctuation index. The greater the video data transmission quality coefficient fluctuation value in a period, the more unstable the video data transmission quality, and the video picture may appear to be stuck, blurred, etc. This will directly lead to the increase of the audio and video fluctuation index. The greater the audio synchronization coefficient fluctuation value in a period, the worse the synchronization between audio and video, and the audio may lag, lead or be interrupted, which will seriously affect the audio and video experience and lead to the increase of the audio and video fluctuation prediction index.

[0077] Specifically, whether to adjust the audio and video fluctuation stability of the target data is determined, and the specific determination process is as follows: comparing the audio and video fluctuation index with the audio and video fluctuation index threshold value; if the audio and video fluctuation index is less than or equal to the audio and video fluctuation index threshold value, it is determined that the initial running parameter is not adjusted, and the audio and video fluctuation stability of the target data is not adjusted; if the audio and video fluctuation index is greater than the audio and video fluctuation index threshold value, it is determined that the audio and video fluctuation stability of the target data is adjusted, and the specific adjustment process is as follows: based on the audio and video fluctuation index and the audio and video fluctuation index threshold value, an audio and video fluctuation index deviation value is obtained, and the encoding rate is reduced based on the audio and video fluctuation index deviation value, so as to reduce the transmission delay and thus reduce the audio and video fluctuation index.

[0078] It should be explained that the above-mentioned audio and video fluctuation index threshold value represents the minimum audio and video fluctuation index value allowed in the video database, and the above-mentioned initialization operating parameters refer to the platform operating parameters corresponding to the audio and video fluctuation index (such as the network packet disorder rate) being input into the video database as the basic value; the above-mentioned audio and video fluctuation index deviation value refers to the difference between the audio and video fluctuation index and the audio and video fluctuation index threshold value; the above-mentioned reduction adjustment of the coding rate based on the audio and video fluctuation index deviation value is specifically carried out as follows: the coding rate reduction coefficient corresponding to each audio and video fluctuation index deviation value interval is stored in the video database, and the obtained audio and video fluctuation index deviation value is input into the video database. The video database can match the corresponding audio and video fluctuation index deviation value interval, and the reduction coefficient corresponding to the interval is the coefficient by which the coding rate needs to be adjusted. The obtained reduction coefficient is multiplied by the original coding rate, and the result obtained is the coding rate that needs to be reduced; the above-mentioned coding rate reduction coefficient is less than 1, which indicates the proportion by which the coding rate needs to be reduced.

[0079] In a specific embodiment, during the data transmission and processing process, audio and video signals are easily affected by various factors and fluctuate. These fluctuations may affect the accurate judgment of students' mental health. By obtaining and processing the fluctuation parameters, the video image is more stable and the audio sound is clearer, reducing misjudgments caused by signal fluctuations, further improving the quality and reliability of the data, and thus improving the accuracy of students' mental health assessment.

[0080] Step 4: Automatically extract the characteristic information of the adjusted target data, thereby obtaining the characteristic information of the students' mental health data.

[0081] The above-mentioned feature information refers to the contour information in the video data.

[0082] In a specific embodiment, the present invention provides a method for dynamic assessment of student mental health based on deep learning, obtains comprehensive mental health data from the target application platform, monitors the processing process in real time, and dynamically analyzes video transmission quality parameters with the help of deep learning to ensure the complete and accurate transmission of video information; accurately collects and evaluates audio synchronization parameters to achieve precise matching of audio and video information; deeply processes audio and video fluctuation parameters to reduce data interference and improve data quality; at the same time, automatically extracts adjusted data features and outputs highly reliable student mental health information. The entire process is automated and intelligent, effectively eliminating interference factors and accurately capturing key information, providing a reliable basis for student mental health assessment and significantly improving the accuracy of assessment results.

[0083] Reference Figure 2As shown, first, the system calculates the video quality coefficient and compares it with the preset threshold; if the coefficient meets the standard, it indicates that the current video data transmission quality is good, and the audio synchronization coefficient calculation link can be directly entered; if it does not meet the standard, the system will optimize and adjust the video buffer. After the adjustment is completed, the improvement rate of the video data transmission quality coefficient is measured; based on the comparison result of the improvement rate with the limit value, if the improvement rate meets the standard, it means that the video transmission performance is improved but there may be a risk of delay, and the system will reduce the cycle length accordingly; if it does not meet the standard, in order to avoid the problem of freezing, the system will extend the cycle length; refer to Figure 3 As shown, the audio synchronization coefficient is then calculated and compared with the threshold. If it meets the standard, it means that the audio synchronization is in good condition. If it does not meet the standard, the system will first make targeted adjustments to the audio parameters and then monitor the audio synchronization again. If it still does not meet the standard after adjustment, the system will switch to an efficient encoding algorithm to improve the audio synchronization performance. Figure 4 As shown in the figure, the audio-visual fluctuation index is further calculated, and the audio-visual fluctuation situation is judged by comparing it with the threshold. If it meets the standard, the system will initialize the platform operating parameters corresponding to the current audio-visual fluctuation index to provide a benchmark for subsequent evaluation; if it does not meet the standard, it indicates that the audio-visual fluctuation is large, and the system will reduce the bit rate and reduce the transmission delay, thereby reducing the audio-visual fluctuation index; finally, the process reaches the end node, completing the comprehensive and refined evaluation and processing of video, audio data and audio-visual fluctuations in the process of dynamic evaluation of students' mental health, providing reliable data support for accurate judgment of students' mental health status.

[0084] Reference Figure 5 As shown, the second aspect of the present invention provides a student mental health dynamic assessment system based on deep learning, including: a video transmission quality adjustment module, an audio synchronization adjustment module, an audio-visual fluctuation adjustment module, a feature extraction module and a video database.

[0085] The video transmission quality adjustment module is connected to the audio synchronization adjustment module, the audio synchronization adjustment module is connected to the audio-visual fluctuation adjustment module, the audio-visual fluctuation adjustment module is connected to the feature extraction module, and the video transmission quality adjustment module, the audio synchronization adjustment module, the audio-visual fluctuation adjustment module and the feature extraction module are collectively connected to the video database; the video transmission quality adjustment module is used to collect student mental health data through the target application platform, mark it as target data, monitor the processing process of the target data, obtain and analyze video transmission quality parameters, and dynamically judge whether to adjust the video transmission process of the target data based on deep learning; the audio synchronization adjustment module is used to collect and evaluate audio synchronization parameters, and judge whether to adjust the audio synchronization process of the target data; the audio-visual fluctuation adjustment module is used to obtain and process audio-visual fluctuation parameters, and judge whether to adjust the audio-visual fluctuation stability of the target data; the feature extraction module is used to automatically extract the feature information of the adjusted target data, thereby obtaining the feature information of the student mental health data; the video database is used to store parameters involved in a student mental health dynamic assessment system based on deep learning.

[0086] The above content is merely an example and explanation of the structure of the present invention. Those skilled in the art may make various modifications or additions to the described specific embodiments or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the present invention, they should all fall within the scope of protection of the present invention.

Claims

1. A method for dynamic assessment of students' mental health based on deep learning, characterized by: include: Step 1: Collect student mental health data through the target application platform and mark it as target data. Monitor the processing of the target data, obtain and analyze video transmission quality parameters, and dynamically determine whether to adjust the video transmission process of the target data based on deep learning. Step 2: Collect and evaluate audio synchronization parameters, and determine whether to adjust the audio synchronization process of the target data; Step 3: Obtain and process audio-visual fluctuation parameters to determine whether to adjust the audio-visual fluctuation stability of the target data; Step 4: Automatically extract the characteristic information of the adjusted target data, thereby obtaining the characteristic information of the students' mental health data.

2. The method for dynamic assessment of student mental health based on deep learning according to claim 1, characterized in that: The acquisition and analysis of video transmission quality parameters, the specific analysis process is as follows: The video transmission quality parameters include the peak-to-average power ratio of the video transmission channel, the channel fading depth factor of the video transmission channel, and the inter-symbol interference power factor of the video transmission channel; By introducing influence coefficients, we quantify the influence of the peak-to-average power ratio of the video transmission channel on the video data transmission quality coefficient, the influence of the channel fading depth factor of the video transmission channel on the video data transmission quality coefficient, and the influence of the inter-symbol interference power factor of the video transmission channel on the video data transmission quality coefficient. By coupling each influence degree, we can obtain the video data transmission quality coefficient, where the video data transmission quality coefficient represents the quality of the video data during the transmission process.

3. The method for dynamic assessment of student mental health based on deep learning according to claim 1, characterized in that: The specific process of determining whether to adjust the video transmission process of the target data is as follows: comparing the video data transmission quality coefficient with the video data transmission quality threshold; If the video data transmission quality coefficient is greater than or equal to the video data transmission quality threshold, it is determined that the video transmission process of the target data is not adjusted; If the video data transmission quality coefficient is less than the video data transmission quality threshold, it is determined that the video transmission process of the target data is adjusted.

4. The method for dynamic assessment of student mental health based on deep learning according to claim 3 is characterized by: The video transmission process of the target data is adjusted, and the specific adjustment process is as follows: Based on the video data transmission quality coefficient and the video data transmission quality threshold, the video data transmission quality coefficient deviation value is obtained, and the increase coefficient is matched based on the video data transmission quality coefficient deviation value, so as to increase the buffer area; and the video data transmission quality coefficient improvement rate within the adjustment period is monitored, and compared with the video data transmission quality coefficient improvement rate limit value to obtain a comparison result, and the duration corresponding to the adjustment period is adjusted according to the comparison result.

5. The method for dynamic assessment of student mental health based on deep learning according to claim 1, characterized in that: The specific evaluation process of collecting and evaluating audio synchronization parameters is as follows: The audio synchronization parameters include a zero-crossing rate difference of the audio signal, an information entropy difference of the audio signal, and a spectral kurtosis difference of the audio signal; By introducing influence coefficients, the influence of the zero-crossing rate difference of the audio signal on the audio synchronization coefficient, the influence of the information entropy difference of the audio signal on the audio synchronization coefficient, the influence of the spectral kurtosis difference of the audio signal on the audio synchronization coefficient, and the influence of the video data transmission quality coefficient on the audio synchronization coefficient are quantified respectively. The influence degrees are coupled to obtain the audio synchronization coefficient, where the audio synchronization coefficient represents the degree of synchronization between audio and video.

6. The method for dynamic assessment of student mental health based on deep learning according to claim 1, characterized in that: The specific process of determining whether to adjust the audio synchronization process of the target data is as follows: The audio synchronization coefficient is compared with the audio synchronization coefficient threshold. If the audio synchronization coefficient is greater than or equal to the audio synchronization coefficient threshold, it is determined that the audio synchronization process of the target data is not adjusted; if the audio synchronization coefficient is less than the audio synchronization coefficient threshold, it is determined that the audio synchronization process of the target data is adjusted.

7. The method for dynamic assessment of student mental health based on deep learning according to claim 6, characterized in that: The audio synchronization process of the target data is adjusted, and the specific adjustment process is: Based on the audio synchronization threshold and the audio synchronization coefficient, an audio synchronization coefficient deviation value is obtained, and the audio synchronization coefficient deviation value is compared with the audio synchronization coefficient deviation threshold; If the audio synchronization coefficient deviation value is greater than or equal to the audio synchronization coefficient deviation threshold, the audio sampling rate is increased based on the audio synchronization coefficient deviation value, a secondary audio synchronization coefficient is obtained, and the secondary audio synchronization coefficient is compared with the audio synchronization system threshold. If the secondary audio synchronization coefficient is still less than the audio synchronization system threshold, a secondary adjustment is performed; if the secondary audio synchronization coefficient is greater than or equal to the audio synchronization system threshold, no secondary adjustment is performed. If the audio synchronization coefficient deviation value is less than the audio synchronization coefficient deviation threshold, the audio frame interval duration is reduced based on the audio synchronization coefficient deviation value.

8. The method for dynamic assessment of student mental health based on deep learning according to claim 1, characterized in that: The specific process of obtaining and processing the audio-visual fluctuation parameters is as follows: The audio-visual fluctuation parameters include the target data network packet disorder rate factor and the target data fractal dimension factor; By introducing influence coefficients, we quantify the influence of the target data network packet disorder rate factor on the audio-visual fluctuation index, the influence of the target data fractal dimension factor on the audio-visual fluctuation index, the influence of the video data transmission quality coefficient fluctuation value within a period on the audio-visual fluctuation index, and the influence of the audio synchronization coefficient fluctuation value within a period on the audio-visual fluctuation index. By coupling each influence degree, we can obtain the audio-visual fluctuation index, where the audio-visual fluctuation index represents the fluctuation of the synchronization between video and audio.

9. The method for dynamic assessment of student mental health based on deep learning according to claim 1, characterized in that: The specific process of determining whether to adjust the stability of the audio-visual fluctuation of the target data is as follows: The audio-visual fluctuation index is compared with the audio-visual fluctuation index threshold; if the audio-visual fluctuation index is less than or equal to the audio-visual fluctuation index threshold, the operating parameters are initialized and the audio-visual fluctuation stability of the target data is not adjusted; if the audio-visual fluctuation index is greater than the audio-visual fluctuation index threshold, the audio-visual fluctuation stability of the target data is adjusted. The specific adjustment process is: based on the audio-visual fluctuation index and the audio-visual fluctuation index threshold, the audio-visual fluctuation index deviation value is obtained, and the coding rate is reduced based on the audio-visual fluctuation index deviation value.

10. A system using the deep learning-based student mental health dynamic assessment method according to any one of claims 1 to 9, characterized in that: include: The video transmission quality adjustment module is used to collect student mental health data through the target application platform, mark it as target data, monitor the processing of the target data, obtain and analyze video transmission quality parameters, and dynamically determine whether to adjust the video transmission process of the target data based on deep learning; An audio synchronization adjustment module is used to collect and evaluate audio synchronization parameters and determine whether to adjust the audio synchronization process of the target data; The audio-visual fluctuation adjustment module is used to obtain and process audio-visual fluctuation parameters and determine whether to adjust the audio-visual fluctuation stability of the target data; The feature extraction module is used to automatically extract the feature information of the adjusted target data, thereby obtaining the feature information of the students' mental health data.

Images