A platform database intelligent evaluation system based on big data analysis

Abstract

The present application relates to the technical field of classroom evaluation, especially to a platform database intelligent evaluation system based on big data analysis, the present application collects basic image information and audio information in the teaching space to jump analysis, determines the interactive period and extracts the interactive state characteristics of the monitoring object based on the optimized image information after optimization processing, combines the interactive duration of the monitoring object to calculate the interactive state quality representation value of the monitoring object, calls the audio information of the interactive period, matches with the content of the standard sample, analyzes the interactive matching degree of the monitoring object, combines the interactive state quality representation value to determine whether to mark the key time domain segment, and analyzes the teaching evaluation of the monitoring object in response to the marking result, the present application reflects the learning state through the extraction of focused features, and the interactive data is convenient for the teaching subject to carry out targeted review, realizes the accurate monitoring and evaluation of the teaching process.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of classroom evaluation, in particular to a platform database intelligent evaluation system based on big data analysis. BACKGROUND

[0002] With the acceleration of education modernization, classroom teaching gradually transforms to an interactive and personalized mode, and the demand for teaching quality evaluation also gradually extends from result-oriented to process-oriented, and needs to cover dynamic data and detailed characteristics in the whole teaching process. Under this trend, the evaluation needs of educators for classroom teaching are increasingly refined, not only need to master the overall teaching effect of the class, but also hope to deeply understand the real-time state of teacher-student interaction and the micro-dimension information such as the concentration degree of students in the teaching process, so as to optimize the teaching strategy and improve the classroom efficiency.

[0003] Chinese patent application publication No. CN119809461A discloses a teaching evaluation system based on artificial intelligence and big data, relating to the field of education evaluation, solving the problem of strong subjectivity and single data dimension in the existing teaching evaluation which relies on artificial questionnaire survey or examination results, comprising a data processing module, a basic analysis module, an expansion analysis module and a comprehensive evaluation module. The data processing module is used for processing the average score of each cultural course test of the class, sending the processed score data to the basic analysis module, and sending the processed expansion data to the expansion analysis module. The basic analysis module is used for analyzing the score level and stability of the class, the expansion analysis module is used for analyzing the expansion ability of students, and the result output module is used for generating a comprehensive teaching evaluation of the class teacher. The present application realizes comprehensive evaluation of teaching situation based on artificial intelligence and big data.

[0004] However, the existing technology still has the following problems,

[0005] Traditional classroom evaluation methods mostly rely on artificial observation and recording, post-questionnaire survey or single examination score analysis, which have obvious limitations. Among them, artificial observation is easily affected by subjective experience, and the evaluation results lack objectivity and consistency, and it is difficult to cover the teaching details of the whole period of classroom teaching and multiple dimensions. Questionnaire survey feedback is lagging behind, which cannot capture the dynamic changes in the teaching process in real time, and the sample size is limited, which makes it difficult to reflect the overall teaching effect. The evaluation mode focusing on examination scores focuses on result-oriented, ignoring process indicators such as student interaction and concentration in the classroom, resulting in single evaluation dimension and inability to fully present the real situation of the teaching process. SUMMARY

[0006] To this end, the application provides a platform database intelligent evaluation system based on big data analysis, to overcome the problem that the traditional classroom evaluation method in the prior art relies on manual observation and recording, post-class questionnaire survey or single test score analysis, and has obvious limitations.

[0007] To achieve the above object, the application provides a platform database intelligent evaluation system based on big data analysis, which comprises:

[0008] The acquisition module comprises image acquisition units arranged in a plurality of teaching spaces to obtain basic image information in the teaching spaces and audio acquisition units to obtain audio information;

[0009] The total control module is connected with the acquisition module to obtain the image information acquired by each image acquisition unit, performs jump analysis based on the basic image parameters of each image frame in the image information, including respectively constructing time domain variation curves for the basic image parameters, determining key time domain segments according to the jump features of the corresponding time domain variation curves, and optimizing the basic image information of the key time domain segments;

[0010] The interactive analysis module is connected with the total control module to determine an interactive period and extract interactive state features of the monitoring object based on the optimized image information after the optimization, calculate the interactive state quality representation value of the monitoring object in combination with the interactive duration of the monitoring object;

[0011] The matching evaluation module is connected with the interactive analysis module to call the audio information of the interactive period, match with the content of a standard sample, analyze the interactive matching degree of the monitoring object, and determine whether to mark the key time domain segments in combination with the interactive state quality representation value;

[0012] The response module is connected with the matching evaluation module, responds to the marking result of the matching evaluation module, analyzes the teaching evaluation, locks the pre-sequence time domain segment and the post-sequence time domain segment of the key time domain segment, identifies the concentration features of the monitoring object, evaluates the concentration degree representation parameter of the monitoring object, analyzes whether the teaching concentration requirement is met, and determines whether to upload the optimized image information and the audio information of the corresponding time domain segment to the review database;

[0013] The interactive state features include the number of mouth movement pauses and the uniformity of pause intervals, and the concentration features include the eye following switching speed and the eye concentration duration variation amplitude.

[0014] Further, the total control module is used to determine the key time domain segments, comprising:

[0015] calling a time domain variation curve for a basic image parameter, including a chroma value time domain variation curve and a luminance time domain variation curve;

[0016] If any time domain variation curve has a time domain segment meeting a jump condition, the time domain segment is determined as a key time domain segment;

[0017] The jump condition includes a variance of peak values at each time in the curve segment being greater than a variance threshold.

[0018] Further, the interaction analysis module is configured to determine an interaction time period, including:

[0019] calling audio information to identify speaking audio corresponding to the teaching subject and the monitoring object;

[0020] determining a questioning interaction triggering time of the teaching subject and a speaking ending time of the monitoring object;

[0021] determining a time period between the questioning interaction triggering time and the speaking ending time as the interaction time period.

[0022] Further, the interaction analysis module is configured to calculate an interaction state quality representation value of the monitoring object, including:

[0023] taking a ratio of the number of mouth movement pauses to a number of mouth movement pause threshold as a first interaction state feature;

[0024] taking a ratio of pause interval uniformity to a pause interval uniformity threshold as a second interaction state feature;

[0025] taking a sum of the first interaction state feature and the second interaction state feature as the interaction state quality representation value.

[0026] Further, the matching evaluation module is configured to analyze an interaction matching degree of the monitoring object, including:

[0027] converting a standard sample and audio information into sample text information and audio text information;

[0028] identifying a plurality of keywords corresponding to the sample text information and the audio text information;

[0029] taking a cosine similarity between the keywords as the interaction matching degree.

[0030] Further, the matching evaluation module is configured to determine whether to mark the key time domain segment, including:

[0031] if the monitoring object in the key time domain segment does not meet an interaction state condition, marking the key time domain segment;

[0032] The interaction state condition comprises that an interaction state quality characteristic value is less than an interaction state quality characteristic threshold value, and an interaction matching degree is greater than an interaction matching degree threshold value.

[0033] Further, the response module is used to analyze the teaching evaluation, and further, the response module is used to evaluate a concentration degree characteristic parameter of the monitoring object, comprising:

[0034] a ratio of the eye following switching speed to an eye following switching speed threshold value is taken as a first concentration degree feature;

[0035] a ratio of an eye concentration time length variation amplitude to an eye concentration time length variation amplitude threshold value is taken as a second concentration degree feature;

[0036] a sum of the first concentration degree feature and the second concentration degree feature is taken as the concentration degree characteristic parameter.

[0037] Further, the response module is used to analyze whether a teaching concentration requirement is met, comprising:

[0038] If the concentration degree characteristic parameter of the detection target is greater than or equal to a concentration degree characteristic parameter threshold value, the teaching concentration requirement is met.

[0039] Further, the response module is used to determine whether to upload the optimized image information and the audio information of the corresponding time domain segment to a review database, comprising:

[0040] If the teaching concentration condition is met, it is determined to upload the optimized image information and the audio information of the corresponding time domain segment to the review database.

[0041] Compared with the prior art, the present application sets a collection module, which comprises image collection units arranged in a plurality of teaching spaces to obtain basic image information in the teaching spaces and audio collection units to obtain audio information; a general control module is used to obtain image information obtained by each image collection unit, and to perform jump analysis based on basic image parameters of each image frame in the image information; an interaction analysis module is used to determine an interaction period and extract an interaction state feature of the monitoring object based on the optimized image information after optimization processing, and to calculate an interaction state quality characteristic value of the monitoring object in combination with an interaction duration of the monitoring object; a matching evaluation module is used to call audio information of the interaction period, match with the content of a standard sample, analyze an interaction matching degree of the monitoring object, and determine whether to mark a key time domain segment in combination with the interaction state quality characteristic value; and a response module responds to the marking result of the matching evaluation module and analyzes the teaching evaluation of the monitoring object. The present application reflects the learning state through the extraction of the concentration feature, facilitates targeted review of the teaching subject in combination with the interaction data, and realizes precise monitoring and evaluation of the teaching process.

[0042] Especially, the present application considers that the traditional teaching evaluation often needs to manually view the complete teaching video frame by frame, which is easy to cause ineffective analysis of smooth periods, such as one-way teaching of teachers and quiet listening of students. According to the analysis of the basic image parameters that can be directly obtained from the image information, the period when the teaching state changes significantly, such as the body movement change during the interaction between teachers and students, is automatically recognized, the key time domain segment of the dynamic change of teaching is locked, and then the image information corresponding to the key time domain segment is optimized for the subsequent accurate analysis of the interactive and focused state. For the scene of large-scale classroom teaching evaluation analysis, the algorithm resource can be saved, and the waste of evaluation resources can be avoided.

[0043] Especially, the present application focuses on the interaction mode and participation state between the monitoring object and the teaching subject presented by the image information corresponding to the key time domain segment. In the teaching interaction scene, the number of mouth action pauses of the monitoring object refers to the cumulative number of mouth action stops in the speaking process, which reflects the degree of expression flow. The more the pause times are, the lower the fluency of interactive expression is. The pause interval uniformity refers to the uniformity of the time interval between the mouth action pauses in the speaking process of the monitoring object. In the actual interaction process, high-quality interaction not only needs fluent expression, but also needs coherent thinking to reflect the degree of coherent thinking. Then, the interactive state quality representation value is calculated comprehensively to quantify the effective degree of regional interactive participation while representing the process completeness of interaction. At the same time, the interactive matching degree represents the effectiveness of the interaction content between the teaching subject and the monitoring object, and reflects the result quality of the teaching subject. The combination of the two realizes the double check of formal compliance and content effectiveness, avoids the situation that the form meets the standard but the content deviates, makes the interactive quality judgment more comprehensive and accurate, provides data support for subsequent determination of whether to mark the key time domain segment. In addition, the teaching subject can directly locate when reviewing later, so that the problem tracing does not need to repeatedly watch the video, and the source can be locked through related data to improve the teaching improvement efficiency.

[0044] Especially, the application not only considers the key time domain segment for interaction, but also locks the preceding time domain segment and the subsequent time domain segment of the key time domain segment, and constructs a complete data chain of "pre-interaction focus state-interaction quality-post-interaction state" in combination with the focus features presented by the monitoring object. The eye following switching speed refers to the response speed of the eye line of the monitoring object switching with the change of the core visual elements in the teaching scene, so as to reflect the adaptation degree and active attention willingness of the monitoring object to the teaching rhythm. In the teaching process, the teaching subject will guide the monitoring object to focus on the core information by switching the courseware, moving the teaching tool, pointing to the key content and the like. The eye focus time length change amplitude refers to the fluctuation degree of the listening focus time length of the monitoring object in the continuous teaching period, so as to reflect the stability degree of the attention concentration of the monitoring object. High-quality classroom focus not only needs to be concentrated for a short time, but also needs to be continuously stable. The focus time length change amplitude of the preceding time domain segment-key time domain segment-subsequent time domain segment reflects the focus change of the monitoring object, depicts the dynamic evolution track of the focus of the monitoring object, and evaluates the influence direction and degree of the key time domain segment, such as the interaction period, and measures the direct influence of the interaction period on the focus stability of the monitoring object. Therefore, the application characterizes the effective degree of the focus state of the monitoring object in the listening process by evaluating the focus degree characterization parameter, so as to provide data support for subsequent analysis of whether the teaching focus requirement is met. The focus feature data corresponding to the key time domain segment and the preceding and subsequent time domain segments stored in the review database are used as the verification basis for the teaching improvement effect, so that the teaching subject can carry out targeted review, and realize the accurate monitoring and evaluation of the teaching process. BRIEF DESCRIPTION OF DRAWINGS

[0045] Figure 1 The functional module diagram of the platform database intelligent evaluation system based on big data analysis of the application embodiment;

[0046] Figure 2 The logic determination diagram for determining the key time domain segment of the application embodiment;

[0047] Figure 3 The logic determination diagram for determining whether to mark the key time domain segment of the application embodiment;

[0048] Figure 4 The logic determination diagram for analyzing whether the teaching focus requirement is met of the application embodiment. DETAILED DESCRIPTION

[0049] In order to make the objects and advantages of the application clearer and more apparent, the application will be further described below in combination with embodiments; it should be understood that the specific embodiments described herein are only used to explain the application, and do not limit the application.

[0050] The preferred embodiments of the present application will be described below with reference to the accompanying drawings. Those skilled in the art will understand that the embodiments are only used to explain the technical principles of the present application and are not intended to limit the protection scope of the present application.

[0051] It should be noted that, in the description of the present application, the terms indicating the direction or positional relationship of "upper", "lower", "left", "right", "inner", "outer" and the like are based on the direction or positional relationship shown in the drawings, which is only for the convenience of description, and does not indicate or imply that the device or element must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application.

[0052] In addition, it should be noted that, in the description of the present application, unless otherwise explicitly specified and limited, the terms "mounting", "connecting", "connecting" should be understood broadly, for example, it can be fixedly connected, or it can be detachably connected, or integrally connected; it can be mechanically connected, or it can be electrically connected; it can be directly connected, or it can be indirectly connected through an intermediate medium; it can be the communication inside two elements. Those skilled in the art can understand the specific meaning of the above terms in the present application according to the specific circumstances.

[0053] Please refer to Figure 1 As shown in the figure, it is a functional module diagram of the platform database intelligent evaluation system based on big data analysis of the embodiment of the present application, the platform database intelligent evaluation system based on big data analysis of the embodiment of the present application comprises:

[0054] The acquisition module comprises an image acquisition unit arranged in a plurality of teaching spaces to obtain basic image information in the teaching space and an audio acquisition unit to obtain audio information;

[0055] The total control module is connected with the acquisition module to obtain the image information acquired by each image acquisition unit, and performs jump analysis based on the basic image parameters of each image frame in the image information, including, respectively constructing a time domain variation curve for the basic image parameters, determining a key time domain segment according to the jump feature of the corresponding time domain variation curve, and optimizing the basic image information of the key time domain segment;

[0056] The interactive analysis module is connected with the total control module to determine an interactive period and extract an interactive state feature of the monitoring object based on the optimized image information after the optimization, and calculate the interactive state quality representation value of the monitoring object in combination with the interactive duration of the monitoring object;

[0057] a matching evaluation module connected with the interaction analysis module, configured to call audio information of the interaction period, match the content of the standard sample, analyze interaction matching degree of the monitoring object, combine the interaction state quality representation value, and determine whether to mark the key time domain segment;

[0058] a response module connected with the matching evaluation module, configured to analyze teaching evaluation in response to a marking result of the matching evaluation module, lock a pre-sequence time domain segment and a post-sequence time domain segment of the key time domain segment, identify attention characteristics of the monitoring object, evaluate an attention degree representation parameter of the monitoring object, analyze whether a teaching attention requirement is met, and determine whether to upload optimized image information and audio information of the corresponding time domain segment to a review database.

[0059] The interaction state characteristics include a number of mouth movement pauses and pause interval uniformity, and the attention characteristics include eye following switching speed and eye attention time length variation amplitude.

[0060] Specifically, the basic image information includes jump characteristics, the optimized image information includes interaction state characteristics, interaction duration of the monitoring object, and attention characteristics, and the audio information includes interaction matching degree.

[0061] The specific structure of the image acquisition unit and the audio acquisition unit is not limited, and only needs to have the function of acquiring basic image information and audio information in the teaching space. High-definition cameras are arranged on the walls in the teaching space to acquire actions and facial states of the teaching subject and the monitoring object. Meanwhile, directional microphones are matched to ensure the clarity of the interaction audio, for example, the teaching subject wears a neck clip directional microphone, and the monitoring object region is arranged with an array type desktop microphone. The teaching subject refers to a teacher who teaches, and the monitoring object refers to a student, which will not be repeated here.

[0062] Specifically, please refer to Figure 2 The total control module is configured to determine the key time domain segment, and includes:

[0063] The time domain variation curve for the basic image parameter includes a chroma value time domain variation curve and a brightness time domain variation curve.

[0064] If any time domain variation curve has a time domain segment meeting the jump condition, the time domain segment is determined as the key time domain segment.

[0065] The jump condition includes that a variance of peak values at each time in the curve segment is greater than a variance threshold.

[0066] Specifically, the present application considers that the traditional teaching evaluation often needs to manually view the complete teaching video frame by frame, which is easy to cause ineffective analysis of smooth periods, such as one-way teaching by the teacher and quiet listening by the students. According to the analysis of the basic image parameters that can be directly obtained from the image information, the periods where the teaching state changes significantly, such as the body movement changes during the interaction between the teacher and the students, are automatically recognized, the key time domain segments of the dynamic changes of teaching are locked, and the image information corresponding to the key time domain segments is optimized for subsequent accurate analysis of the interactive and focused states. For the scene of large-scale classroom teaching evaluation and analysis, the algorithm resource can be saved, and the waste of evaluation resources can be avoided.

[0067] In the embodiment, the purpose of setting the variance threshold is to represent the case that the basic image information has a large fluctuation in the corresponding basic image parameters in the time domain segment. By obtaining historical image information of a teaching subject completing several classroom teachings, calling the variance historical data of the peak values of the time domain change curve at each time, solving the mean value of the variance, and taking it as the reference value under normal circumstances, the variance threshold is determined as the product of the mean value of the variance and the variance deviation coefficient, wherein the variance deviation coefficient is selected within the interval [1.1, 1.15].

[0068] Specifically, the time domain change curve for the basic image parameter is constructed by the following method, including:

[0069] A rectangular coordinate system is constructed with time as the horizontal axis and chroma value as the vertical axis;

[0070] The coordinate points of the chroma value at each time are calibrated in the rectangular coordinate system;

[0071] The coordinate points are connected by a smooth curve to obtain the chroma value time domain change curve.

[0072] Correspondingly, a rectangular coordinate system is constructed with time as the horizontal axis and brightness as the vertical axis;

[0073] The coordinate points of the brightness at each time are calibrated in the rectangular coordinate system;

[0074] The coordinate points are connected by a smooth curve to obtain the brightness time domain change curve.

[0075] Specifically, the way of constructing the time domain change curve for the basic image parameter is not limited, for example, the time domain curve can be fitted by Matlab related fitting software, which will not be repeated here.

[0076] Specifically, the interaction analysis module is used to determine the interaction period, including:

[0077] Audio information is called to identify the speaking audio of the teaching subject and the monitoring object;

[0078] determine the question interaction trigger moment of the teaching subject and the speech end moment of the monitoring object;

[0079] determine the period between the question interaction trigger moment and the speech end moment as the interaction period.

[0080] Specifically, for the determination of the question interaction trigger moment, a pre-trained semantic model, such as BERT, can be used to determine whether the speech of the teaching subject contains a question guide sentence, which is a prior art and will not be described in detail.

[0081] Specifically, the interaction analysis module is configured to calculate the interaction state quality representation value of the monitoring object, including:

[0082] determine the ratio of the number of mouth movement pauses to the number of mouth movement pause thresholds as a first interaction state feature;

[0083] determine the ratio of the pause interval uniformity to the pause interval uniformity threshold as a second interaction state feature;

[0084] determine the sum of the first interaction state feature and the second interaction state feature as the interaction state quality representation value.

[0085] In this embodiment, the purpose of setting the number of mouth movement pause thresholds and the pause interval uniformity threshold is to represent the situation that the expression of the monitoring object is not smooth enough, and the thought is not clear enough and the logic is not complete enough in the interaction process with the teaching subject. By obtaining historical image information of the same teaching subject completing several classroom teaching, calling the historical data of the number of mouth movement pauses of the monitoring object and the corresponding pause interval uniformity historical data, solving the mean value of the number of mouth movement pauses and the mean value of the pause interval uniformity, and corresponding as the reference value under normal circumstances, the purpose of setting the above three thresholds is to determine the number of mouth movement pause thresholds as the product of the mean value of the number of mouth movement pauses and the first deviation coefficient, and the pause interval uniformity threshold is determined as the product of the mean value of the pause interval uniformity and the second deviation coefficient, wherein the first deviation coefficient is selected within the interval [1.2, 1.4], and the implementation is preferably 1.2, and the second deviation coefficient is selected within the interval [0.9, 0.95], and the implementation is preferably 0.9;

[0086] Wherein, the uniformity of the pause interval is quantified by the standard deviation, and the smaller the standard deviation, the more uniform the pause interval.

[0087] Specifically, the matching evaluation module is configured to analyze the interaction matching degree of the monitoring object, including:

[0088] determine the standard sample and the audio information into sample text information and audio text information;

[0089] identifying a plurality of keywords corresponding to the sample text information and the audio text information;

[0090] using the cosine similarity between the keywords as the interaction matching degree.

[0091] In the embodiment, the expression of the teaching subject in the teaching process is taken as the standard sample, and then the answer information corresponding to the question of the teaching subject in the standard sample and the response information corresponding to the response of the monitored object in the audio information are extracted, and the keywords in the answer information and the response information are matched, which will not be described herein.

[0092] Specifically, referring to FIG. 6, which is a logic determination diagram for determining whether to mark a key time domain segment, the matching evaluation module is used to determine whether to mark the key time domain segment, and includes: Figure 3 If the monitored object in the key time domain segment does not meet the interaction state condition, the key time domain segment is marked;

[0093] If the monitored object in the key time domain segment meets the interaction state condition, the key time domain segment does not need to be marked;

[0094]

[0095] The interaction state condition includes that the interaction state quality representation value is less than the interaction state quality representation threshold value, and the interaction matching degree is greater than the interaction matching degree threshold value.

[0096] In the embodiment, the purpose of setting the interaction matching degree threshold value is to represent the case that the interaction between the teaching subject and the monitored object is effective, the historical audio information of the same teaching subject completing a plurality of classroom teaching is obtained, the historical data of the interaction matching degree is called, the mean value of the interaction matching degree is solved, and is taken as a reference value in a normal case, and based on the purpose of setting the interaction matching degree threshold value, the interaction matching degree threshold value is determined as the product of the mean value of the interaction matching degree and a matching deviation coefficient, wherein the matching deviation coefficient is selected in the interval [1.2, 1.3], and in the implementation, 1.2 is preferred.

[0097] The interaction state quality representation threshold value is determined as the interaction state quality representation value calculated under the condition that the number of mouth movement pauses is equal to the number of mouth movement pause thresholds, and the uniformity of the pause interval is equal to the uniformity of the pause interval threshold.

[0098] ​Specifically, the application focuses on the interaction mode and participation state between the monitoring object and the teaching subject presented by the image information corresponding to the key time domain segment. In the teaching interaction scene, the number of mouth action pauses of the monitoring object refers to the cumulative number of mouth action stops during the speech process, reflecting the degree of expression flow. If the student is familiar with the interactive content, such as knowledge points and problem answers, and speaks coherently, the mouth action will present a continuous and stable state, and the pause frequency will be low. On the contrary, if the student is not familiar with the content, such as not understanding the problem or fuzzy memory of knowledge points, and needs to think frequently during speech, the mouth action will frequently pause, possibly pausing once every 3-5 seconds when answering, or even appearing to be stuck for a long time. Based on this, the more the pause frequency, the lower the fluency of the interactive expression. The uniformity of the pause interval refers to the uniformity of the time interval between the mouth action pauses of the monitoring object during the speech process. In actual interaction process, high-quality interaction not only requires fluent expression, but also requires coherent thinking to reflect the degree of coherent thinking. If the student's thinking is clear and logical in the interaction, the pause interval will present a uniform feature. If the student's thinking is chaotic and the logic jumps, the pause interval will fluctuate long and short. Then, the interactive state quality representation value is calculated comprehensively to quantify the effective degree of regional interaction participation and represent the process completeness of the interaction. At the same time, the interaction matching degree represents the effectiveness of the interaction content between the teaching subject and the monitoring object, for example, whether the answer content of the monitoring object is around the teaching focus and whether it meets the knowledge goal, etc., reflecting the result quality of the teaching subject. The combination of the two realizes the double check of form qualification and content effectiveness, avoids one-sided evaluation such as students actively speaking but answering irrelevant questions, or content being correct but state being passive, such as students answering accurately but pausing frequently and having low participation willingness, so that the interaction quality judgment is more comprehensive and accurate. Data support is provided for subsequent determination of whether to mark the key time domain segment. In addition, the teaching subject can directly locate when reviewing later, so that problem tracing does not need to repeatedly watch the video, and the source can be locked through related data, improving the efficiency of teaching improvement.

[0099] Specifically, the response module analyzes the teaching evaluation in response to the key time domain segment being marked. Specifically, the response module is used to evaluate the concentration degree representation parameter of the monitoring object, including:

[0100] The ratio of the eye following switching speed to the eye following switching speed threshold value is used as the first concentration degree feature;

[0101] The ratio of the eye concentration duration change amplitude to the eye concentration duration change amplitude threshold value is used as the second concentration degree feature;

[0102] The sum of the first concentration degree feature and the second concentration degree feature is used as the concentration degree representation parameter.

[0103] In this embodiment, the purpose of setting the eye following switching speed threshold value and the eye focus duration change amplitude threshold value is to represent the case that the listening focus degree of the monitored object is high. By obtaining historical image information of a same teaching subject completing several times of classroom teaching, calling the historical data of the eye following switching speed and the eye focus duration change amplitude of the monitored object, solving the mean value of the eye following switching speed and the mean value of the eye focus duration change amplitude, and corresponding as the reference value under normal circumstances, based on the purpose of setting the above two threshold values, the eye following switching speed threshold value is determined as the product of the mean value of the eye following switching speed and a speed deviation coefficient, and the eye focus duration change amplitude threshold value is determined as the product of the mean value of the eye focus duration change amplitude and an amplitude deviation coefficient, wherein the speed deviation coefficient is selected within the interval [1.2, 1.25], and in the implementation, 1.2 is preferred, and the amplitude deviation coefficient is selected within the interval [1.15, 1.2], and in the implementation, 1.15 is preferred.

[0104] Specifically, the application not only considers the key time domain segment for interaction, but also locks the preceding time domain segment and the subsequent time domain segment of the key time domain segment, and constructs a complete data chain of "pre-interaction focus state-interaction quality-post-interaction state" in combination with the focus features presented by the monitoring object. The eye following switching speed refers to the response speed of the eye line of the monitoring object to the core visual elements in the teaching scene, such as the teacher's gesture pointing, the page switching of the courseware, and the position change switching of the writing on the blackboard, to reflect the adaptation degree and the active attention willingness of the monitoring object to the teaching rhythm. In the teaching process, the teaching subject will guide the monitoring object to focus on the core information by switching the courseware, moving the teaching tools, and pointing to the key content. If the student is focused on the classroom, the eyes will quickly follow the switching of these visual elements, and the eye following switching speed is fast, which indicates that the student can keep up with the teaching rhythm and does not lag behind. On the contrary, if the student's attention is scattered, the response of the eyes to the switching of the visual elements will be obviously delayed, and the switching speed will be slow, which indicates that the student is out of sync with the teaching rhythm and does not effectively receive the classroom information. The eye focus duration change amplitude refers to the fluctuation degree of the listening focus duration of the monitoring object in the continuous teaching period, to reflect the stability degree of the attention concentration of the monitoring object. High-quality classroom focus not only requires short-term concentration, but also requires stable and continuous concentration. If the student's attention is stable and continuous, the student will continuously focus on the core elements of teaching for a period of time. The focus duration change amplitude of the preceding time domain segment-key time domain segment-subsequent time domain segment reflects the change of the concentration of the monitoring object, depicts the dynamic evolution trajectory of the concentration of the monitoring object, and evaluates the influence direction and degree of the key time domain segment, such as the interaction period, and measures the direct influence of the interaction period on the stability of the concentration of the monitoring object. Therefore, the application characterizes the effective degree of the focus state of the monitoring object in the listening process by evaluating the focus degree characterization parameter, to provide data support for subsequent analysis of whether the teaching focus requirement is met. The focus feature data corresponding to the key time domain segment and the preceding and subsequent time domain segments stored in the review database are used as the verification basis for the teaching improvement effect, to facilitate the teaching subject to conduct targeted review and realize the precise monitoring and evaluation of the teaching process.

[0105] Specifically, please refer to Figure 4 as shown in the figure, which is the logic judgment diagram of the application embodiment for analyzing whether the teaching focus requirement is met. The response module is used to analyze whether the teaching focus requirement is met, which includes:

[0106] If the focus degree characterization parameter of the detection target is greater than or equal to the focus degree characterization parameter threshold, the teaching focus requirement is met.

[0107] If the focus degree characterization parameter of the detection target is less than the focus degree characterization parameter threshold, the teaching focus requirement is not met.

[0108] The concentration degree representation parameter threshold value is predetermined, the eye following switching speed is equal to the eye following switching speed threshold value, and the eye concentration duration change amplitude is equal to the eye concentration duration change amplitude threshold value, and the concentration degree representation parameter calculated is determined as the concentration degree representation parameter threshold value.

[0109] Specifically, the response module is used to determine whether to upload the optimized image information and the audio information of the corresponding time domain segment to the review database, comprising:

[0110] If the teaching concentration condition is met, it is determined to upload the optimized image information and the audio information of the corresponding time domain segment to the review database.

[0111] So far, the technical solutions of the present application have been described in combination with the preferred embodiments shown in the drawings, but those skilled in the art can easily understand that the protection scope of the present application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to the related technical features without departing from the principles of the present application, and the technical solutions after these changes or replacements will fall within the protection scope of the present application.

Claims

1. A platform database intelligent evaluation system based on big data analysis, characterized in that, The method comprises the following steps: a collection module comprising image collection units arranged in a plurality of teaching spaces to obtain basic image information in the teaching spaces and audio collection units to obtain audio information; a total control module configured to obtain image information collected by each image collection unit, perform jump analysis based on basic image parameters of each image frame in the image information, including respectively constructing time domain variation curves for the basic image parameters, determining key time domain segments according to jump features of the corresponding time domain variation curves, and performing optimization processing on the basic image information of the key time domain segments; an interaction analysis module configured to determine an interaction period and extract interaction state features of a monitoring object based on the optimized image information after the optimization processing, calculate an interaction state quality representation value of the monitoring object in combination with an interaction duration of the monitoring object; a matching evaluation module configured to call audio information of the interaction period, match the audio information with content of a standard sample, analyze an interaction matching degree of the monitoring object, and determine whether to mark the key time domain segments in combination with the interaction state quality representation value; a response module configured to analyze teaching evaluation in response to a marking result of the matching evaluation module, lock a preceding time domain segment and a subsequent time domain segment of the key time domain segments, identify concentration features of the monitoring object, evaluate a concentration degree representation parameter of the monitoring object, analyze whether a teaching concentration requirement is met, and determine whether to upload the optimized image information and the audio information of the corresponding time domain segments to a review database; wherein the interaction state features include a number of mouth movement pauses and pause interval uniformity, and the concentration features include eye following switching speed and eye concentration duration variation amplitude; the interaction analysis module is configured to calculate the interaction state quality representation value of the monitoring object, including: using a ratio of the number of mouth movement pauses to a mouth movement pause number threshold value as a first interaction state feature; using a ratio of the pause interval uniformity to a pause interval uniformity threshold value as a second interaction state feature; using a sum of the first interaction state feature and the second interaction state feature as the interaction state quality representation value. 2.The big data analysis based platform database intelligent evaluation system according to claim 1, characterized in that, the total control module is configured to determine the key time domain segments, including: calling time domain variation curves for the basic image parameters, including a chroma value time domain variation curve and a brightness time domain variation curve; if any time domain variation curve has a time domain segment meeting a jump condition, the time domain segment is determined as a key time domain segment; wherein the jump condition includes a variance of peak values at each time in a curve segment being greater than a variance threshold value. 3.The big data analysis based platform database intelligent evaluation system according to claim 1, characterized in that, the interaction analysis module is configured to determine the interaction period, including: calling audio information to identify speaking audio corresponding to a teaching subject and the monitoring object; determining a questioning interaction trigger time of the teaching subject and an end time of the monitoring object's speaking; determining a period between the questioning interaction trigger time and the end time of the speaking as the interaction period. 4.The big data analysis based platform database intelligent evaluation system according to claim 1, characterized in that, the matching evaluation module is configured to analyze the interaction matching degree of the monitoring object, including: converting the standard sample and the audio information into sample text information and audio text information, respectively; A plurality of keywords corresponding to the sample text information and the audio text information are identified; A cosine similarity between the keywords is used as the interaction matching degree. 5.The big data analysis based platform database intelligent evaluation system according to claim 1, characterized in that, The matching evaluation module is configured to determine whether to mark the key time domain segment, including: If the monitored object in the key time domain segment does not satisfy the interaction state condition, the key time domain segment is marked; The interaction state condition includes that an interaction state quality representation value is less than an interaction state quality representation threshold value, and an interaction matching degree is greater than an interaction matching degree threshold value. 6.The big data analysis based platform database intelligent evaluation system according to claim 5, characterized in that, The response module is configured to analyze the teaching evaluation in response to the key time domain segment being marked. 7.The big data analysis based platform database intelligent evaluation system according to claim 1, characterized in that, The response module is configured to evaluate a concentration degree representation parameter of the monitored object, including: A ratio of an eye following switching speed to an eye following switching speed threshold value is used as a first concentration degree feature; A ratio of an eye concentration time length change amplitude to an eye concentration time length change amplitude threshold value is used as a second concentration degree feature; A sum of the first concentration degree feature and the second concentration degree feature is used as the concentration degree representation parameter. 8.The big data analysis based platform database intelligent evaluation system according to claim 7, characterized in that, The response module is configured to analyze whether a teaching concentration requirement is satisfied, including: If the concentration degree representation parameter of the monitored object is greater than or equal to a concentration degree representation parameter threshold value, the teaching concentration requirement is satisfied. 9.The big data analysis based platform database intelligent evaluation system according to claim 8, characterized in that, The response module is configured to determine whether to upload optimized image information and audio information of a corresponding time domain segment to a review database, including: If the teaching concentration condition is satisfied, it is determined to upload the optimized image information and the audio information of the corresponding time domain segment to the review database.

Images