CN116362587A - College classroom teaching evaluation method and system based on artificial intelligence - Google Patents
College classroom teaching evaluation method and system based on artificial intelligence Download PDFInfo
- Publication number
- CN116362587A CN116362587A CN202310121252.9A CN202310121252A CN116362587A CN 116362587 A CN116362587 A CN 116362587A CN 202310121252 A CN202310121252 A CN 202310121252A CN 116362587 A CN116362587 A CN 116362587A
- Authority
- CN
- China
- Prior art keywords
- teaching
- index
- student
- score
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000011156 evaluation Methods 0.000 title claims abstract description 87
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 31
- 238000001514 detection method Methods 0.000 claims abstract description 163
- 238000004458 analytical method Methods 0.000 claims abstract description 130
- 238000013441 quality evaluation Methods 0.000 claims abstract description 92
- 238000000034 method Methods 0.000 claims abstract description 53
- 238000000605 extraction Methods 0.000 claims abstract description 41
- 230000008909 emotion recognition Effects 0.000 claims abstract description 33
- 230000004927 fusion Effects 0.000 claims abstract description 33
- 238000012549 training Methods 0.000 claims abstract description 19
- 238000012360 testing method Methods 0.000 claims abstract description 18
- 239000013598 vector Substances 0.000 claims description 27
- 230000000694 effects Effects 0.000 claims description 24
- 238000004364 calculation method Methods 0.000 claims description 14
- 238000012545 processing Methods 0.000 claims description 11
- 238000010276 construction Methods 0.000 claims description 6
- 230000007935 neutral effect Effects 0.000 claims description 5
- 238000004519 manufacturing process Methods 0.000 claims description 4
- 230000008569 process Effects 0.000 abstract description 17
- 238000005516 engineering process Methods 0.000 abstract description 12
- 230000008451 emotion Effects 0.000 description 16
- 238000011835 investigation Methods 0.000 description 14
- 230000006870 function Effects 0.000 description 7
- 230000014509 gene expression Effects 0.000 description 7
- 238000010586 diagram Methods 0.000 description 5
- 230000001815 facial effect Effects 0.000 description 5
- 230000008449 language Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 239000000284 extract Substances 0.000 description 3
- 230000036544 posture Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000006399 behavior Effects 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 125000004122 cyclic group Chemical group 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 230000003631 expected effect Effects 0.000 description 2
- 238000005457 optimization Methods 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 230000002411 adverse Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002457 bidirectional effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000008921 facial expression Effects 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000000047 product Substances 0.000 description 1
- 230000002269 spontaneous effect Effects 0.000 description 1
- 239000013589 supplement Substances 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/20—Education
- G06Q50/205—Education administration or guidance
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
- G06V40/161—Detection; Localisation; Normalisation
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Business, Economics & Management (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Strategic Management (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Educational Administration (AREA)
- Evolutionary Computation (AREA)
- Tourism & Hospitality (AREA)
- Medical Informatics (AREA)
- Economics (AREA)
- Databases & Information Systems (AREA)
- Educational Technology (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Human Computer Interaction (AREA)
- Entrepreneurship & Innovation (AREA)
- Development Economics (AREA)
- Operations Research (AREA)
- Primary Health Care (AREA)
- Quality & Reliability (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Game Theory and Decision Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the field of classroom teaching quality evaluation, in particular to an artificial intelligence-based college classroom teaching evaluation method and system. Comprising the following steps: constructing a teaching quality evaluation system; constructing a teaching coordination index analysis model comprising a text feature extraction module, a voice feature extraction module, a video feature extraction module, a feature fusion module and an emotion recognition module; constructing a teaching order index analysis model comprising a face detection module, a human body detection module and a fusion analysis module; respectively training and testing a teaching coordination index analysis model and a teaching order index analysis model; analyzing by using a teaching coordination index analysis model to obtain a teaching coordination index score, and analyzing by using a teaching order index analysis model to obtain a teaching order index score; and carrying out teaching quality evaluation according to the scores of the indexes. According to the invention, the artificial intelligence technology is applied to the teaching quality evaluation, so that the objectivity, scientificity and accuracy of the teaching quality evaluation process and result are effectively improved.
Description
Technical Field
The invention relates to the field of classroom teaching quality evaluation, in particular to an artificial intelligence-based college classroom teaching evaluation method and system.
Background
The teaching evaluation of colleges and universities is based on the teaching of teachers and the learning of students, and aims at improving the teaching ability of the teachers, improving the teaching quality of the classroom, and evaluating the teaching design, process and result development of the classroom. The classroom teaching evaluation work of the colleges and universities in China starts at the early 80 s of the 20 th century, and most colleges and universities in China develop the evaluation work of the classroom teaching at present in different degrees, and different evaluation contents, evaluation standards and evaluation systems are formulated, wherein the classroom teaching evaluation mainly comprises internal and external multi-element evaluation, process and performance evaluation judgment and expert and peer field observation.
However, in the prior art, classroom teaching evaluation is mainly in the form of questionnaire and expert scoring, and has certain subjectivity and limitation. With the rapid popularization of artificial intelligence technology, especially in recent years, picture processing technology, voice processing technology and video processing technology have come into rapid progress, the industry has focused attention on the application of artificial intelligence technology in education, and the application of artificial intelligence technology in aspects of classroom behavior analysis, classroom expression recognition, class arrival rate detection and the like.
Disclosure of Invention
The invention aims to overcome at least one defect (deficiency) of the prior art, and provides a college classroom teaching evaluation method and system based on artificial intelligence.
The technical scheme adopted by the invention is as follows:
in a first aspect, an artificial intelligence-based classroom teaching evaluation method for universities is provided, including:
constructing a teaching quality evaluation system, which comprises constructing teacher teaching indexes and student learning indexes, and constructing teaching coordination indexes and teaching order indexes by taking a teacher and a student as evaluation subjects;
constructing a teaching coordination index analysis model comprising a text feature extraction module, a voice feature extraction module, a video feature extraction module, a feature fusion module and an emotion recognition module;
constructing a teaching order index analysis model comprising a face detection module, a human body detection module and a fusion analysis module;
manufacturing a training set and a testing set to train and test the teaching coordination index analysis model and the teaching order index analysis model respectively, so as to obtain a trained teaching coordination index analysis model and a trained teaching order index analysis model;
Collecting teacher teaching index data and student learning index data, collecting teaching coordination index data comprising text data, voice data and video data, and collecting teaching order index data comprising student image data;
calculating a teacher teaching index score and a student learning index score according to the teacher teaching index data and the student learning index data;
analyzing the teaching coordination index according to the teaching coordination index data by using a trained teaching coordination index analysis model, and analyzing the teaching order index according to the teaching order index data by using a trained teaching order index analysis model to obtain a teaching order index score;
calculating a final teaching quality evaluation score according to the index scores;
and carrying out teaching quality evaluation according to the final teaching quality evaluation score.
The teaching quality evaluation system constructed by the invention comprises teacher teaching indexes and student learning indexes taking a teacher and a student as evaluation subjects. However, the index score using a person as an evaluation subject has a certain subjectivity, and if the teaching quality evaluation is performed only by means of subjective data, the accuracy may be low. Therefore, the teaching quality evaluation system constructed by the invention also comprises teaching coordination indexes and teaching order indexes, a multi-dimensional teaching quality evaluation system is constructed, and the teaching coordination indexes and the teaching order indexes which are ignored in the past are introduced into the teaching quality evaluation system, so that the teaching quality evaluation dimension is more comprehensive. According to the invention, an artificial intelligence technology is applied to teaching quality evaluation, a teaching coordination index analysis model and a teaching order index analysis model are constructed to analyze teaching coordination index data and teaching order index data, objective index scores are obtained as objective data, and subjective data and objective data are combined to carry out teaching quality evaluation, so that the objectivity, scientificity and accuracy of a teaching quality evaluation process and a result can be effectively improved.
Further, the collecting teaching coordination index data including text data, voice data and video data, analyzing the teaching coordination index data by using a trained teaching coordination index analysis model to obtain a teaching coordination index score, specifically including:
the voice data and the video data of teachers and students in the class are collected by using a sound collection device and a video collection device, and text recognition is carried out on the voice data to obtain text data;
encoding the text data by using a text feature extraction module of the teaching coordination index analysis model, and extracting text feature vectors of the encoded text data;
processing the voice data by using a voice characteristic extraction module of the teaching coordination index analysis model, and extracting a voice characteristic vector of the processed voice data;
processing the video data by using a video feature extraction module of the teaching coordination index analysis model, and extracting the processed video feature vector;
feature fusion is carried out on the text feature vector, the voice feature vector and the video feature vector by using a feature fusion module of the teaching coordination index analysis model;
carrying out emotion recognition on the fused features by using an emotion recognition module of the teaching coordination index analysis model to obtain a classification result, wherein the classification result comprises a negative state, a neutral state and a positive state;
And obtaining the teaching coordination index score according to the number of the teacher active states and the number of the student active states identified in a class.
When a human expresses a certain emotion, the emotion is usually expressed in various modes such as language expression, action, voice intonation and the like, a great amount of emotion information is contained in a language text of the human, and the emotion content is recognized from the text, so that the emotion state of the human can be better understood, and therefore, text characteristics are required to be extracted for emotion recognition; the human voice contains various acoustic mode information such as voice intonation, pause suppression, frequency change and the like besides language information, and the voice information is important data for emotion recognition, so that voice characteristics are required to be extracted for emotion recognition; more than about half of human expression is reflected by the face, which is sufficient to indicate that human facial expression is critical for emotion recognition, and thus video features need to be extracted to analyze the expressive image of the face for emotion recognition. According to the invention, emotion recognition is carried out from three dimensions of text features, voice features and video features, and compared with emotion recognition carried out by only one feature, the emotion recognition is carried out after feature fusion of the three features, so that the obtained result is more accurate.
Further, the method for obtaining the teaching coordination index score according to the number of the teacher active state and the number of the student active state identified in a class specifically includes:
setting the time of each class as T, and respectively carrying out K emotion recognition on teachers and students in the time T, wherein the number of times that the teachers recognize the positive state is TCount pos The number of times that the student recognizes the positive state is SCount pos The teaching coordination index score is: score A3 ＝(TCount pos +SCount pos )/2K。
In the classroom teaching process, the teaching emotion of a teacher is often transferred to students through expressions, languages and actions, the positive emotion of full enthusiasm can be more mobilized to play a role in learning enthusiasm of the students, and the proper use of the negative emotion can also play a role in frightening discipline and correcting improper behaviors. The learning emotion of the student can be transmitted to a teacher through the expression and the action, the positive emotion shows that the learning enthusiasm of the student is high, the teaching mode of the teacher is suitable, when the student is in a concentrated and input learning state along with the positive emotion in the classroom learning process, the student is more favorable for producing a high-quality learning effect, and the negative learning emotion also reflects that the learning enthusiasm of the student is low, so that the learning quality is influenced. The influence of the emotion expressions of teachers and students on the classroom on the teaching quality is great, so that emotion recognition analysis is carried out through a teaching coordination index analysis model to obtain a teaching coordination index score, teaching quality evaluation is carried out, and the rationality and scientificity of the teaching quality evaluation can be effectively improved.
Further, the collecting includes face image data and teaching order index data of the human body image data, and the analyzing module of the trained teaching order index is used for analyzing and obtaining the teaching order index score according to the teaching order index data, which specifically includes:
using a plurality of image acquisition devices with different angles to acquire student image data in different time periods in a class;
a face detection module using a teaching order index analysis model recognizes face detection frame data in an image according to student image data to obtain a student face detection frame set;
the human body detection module using the teaching order index analysis model recognizes human body detection frame data in the image according to the student image data to obtain a student human body detection frame set;
the method comprises the steps that a fusion analysis module of a teaching order index analysis model is used for fusing a student face detection frame set and a student human body detection frame set to obtain a student target detection frame set, and class arrival rates of different time periods in a class are calculated according to the student target detection frame set;
and calculating according to the class arrival rates of all time periods in a class to obtain the teaching order index score.
The core targets of class rate calculation of students in a class are the students and the faces of the students, but because the class scene is bigger, interference objects such as tables and chairs influence human body detection, and the head postures of the students are various during the class, and complete face information is sometimes not observed. Therefore, when the students are positioned, the expected effect of the single face detection or the single body detection on the seriously shielded target is difficult to achieve. Therefore, the invention provides the teaching order index analysis model for fusing the face detection and the human body detection aiming at complex situations that the intensive distribution and shielding of students in a class can influence the recognition effect and the like, thereby effectively improving the accuracy of the class-to-class rate detection, and the model has the characteristics of high robustness and high accuracy.
Further, the fusion module that uses the teaching order index analysis model fuses student's human face detection frame collection and student's human body detection frame collection, acquires student's detection frame target set, use fusion module to fuse student's human face detection frame collection and student's human body detection frame collection, acquire student's detection frame target set, calculate the class rate of going to in different time slots on the classroom according to classroom student's detection frame target set, specifically include:
let the face detection frame set of student be F= { (x) 1 ,y 1 ,w 1 ,h 1 ),…,(x j ,y j ,w j ,h j ),…,(x m ,y m ,w m ,h m ) X, where x j And y j X-axis coordinates and Y-axis coordinates of a center point of a face detection frame of a jth student, w j The width of the face detection frame of the jth student is h j The height of the face detection frame of the jth student is the total number of faces of the student identified by the face detection module;
let the student human body detection frame set be B = { (x) 1 ,y 1 ,w 1 ,h 1 ),…,(x i ,y i ,w i ,h i ),…,(x n ,y n ,w n ,h n ) X, where x i And y i X-axis coordinates and Y-axis coordinates of the center point of the human body detection frame of the ith student, w i Broadband for human body detection frame of ith student, h i N is the height of the human body detection frame of the ith student, and n is the total human body number of the students identified by the human body detection module;
initializing a student target detection frame set as P, and setting P=F, wherein the element number of the set P is m;
Selecting the kth student from the human body detection frame set B, and then B (k) = (x) k ,y k ,w k ,h k )，x k Representing X-axis coordinate, y of human body detection frame of kth student k Represents the Y-axis coordinate, w, of the human body detection frame of the kth student k The width of the human body detection frame of the kth student is represented, h k Representing the height of a human body detection frame of a kth student;
judging whether the human body detection frame of the kth student contains the human face detection frame of any student in the student human face detection frame set F or intersects with the human face detection frame of any student in the student human face detection frame set F, if not, adding B (k) into P;
traversing the human body detection frame set B to obtain a class student target detection frame set P, and setting the element number of the set P as z, namely the current class student number as z;
setting the number of students in a classroom to be N, setting the current time period to be t, and setting the class arrival rate of the time period t to be:
the teaching order index analysis model constructed by the invention takes face detection as main human body detection as auxiliary calculation class rate, and supplements the face detection through the human body detection, thereby improving the detection effect.
Further, the method for obtaining the teaching order index score according to the class arrival rate of all time periods in a class specifically comprises the following steps:
teaching if the student image data of M time periods are collected in total The order index score is:
wherein Q is M Is the lesson arrival rate of time period M.
In the course of teaching in class, if adverse phenomena such as late arrival, early departure and class escape appear, the teaching order can be seriously influenced, so that the quality of teaching in class is influenced, and therefore, the class arrival rate of students is also a very key ring in a teaching quality evaluation system. The invention uses human face detection and human body detection technology to detect the class arrival rate of students in class in a plurality of different time periods including but not limited to before class, during class, and the like, and calculates the teaching order index score through the class arrival rates of all time periods, thereby carrying out teaching quality evaluation and improving the accuracy of teaching quality evaluation. Meanwhile, the teaching machine can play a certain warning role for students with low class rate, so that class teaching order is maintained, and teaching quality is improved.
Further, the collecting teacher teaching index data and student learning index data specifically includes:
q evaluation results reflecting teaching indexes of teachers in the class are obtained through questionnaires, wherein q is more than 0, and each evaluation result comprises a teaching target index score, a teaching ability index score and a teaching method index score;
And obtaining p evaluation results reflecting the learning index of the students in the class through questionnaires, wherein p is more than 0, and each evaluation result comprises a learning effect index score.
The questionnaire investigation objects of the teacher teaching indexes are teachers and students, wherein teacher evaluation comprises teacher self-evaluation, teaching management department evaluation and teaching quality expert evaluation. The teaching target indexes are the clear degree of teaching ideas of the teacher, the reasonable degree of teaching plan arrangement and the standard reaching degree of course teaching targets; the teaching ability index is the familiarity degree of the teaching content of the investigation teacher and the grasping degree of the weight difficulty of the teaching material; the teaching method index is used for examining the application conditions of various teaching methods of teachers. And the teacher teaching index is evaluated from multiple angles, so that the evaluation result is more comprehensive.
The questionnaire investigation object of the student learning index is a teacher and a student, wherein the student evaluation comprises student self-evaluation and student mutual evaluation, and the learning effect index is the mastering condition of the teaching content of the student classroom. The teacher can not accurately observe the learning conditions of all students in the teaching process, so that the accuracy of the learning index score of the students can be effectively improved by increasing the questionnaire survey of the mutual evaluation of the students.
Further, the calculating the teacher teaching index score and the student learning index score according to the teacher teaching index data and the student learning index data specifically includes:
Combining the teaching target index score, the teaching ability index score and the teaching method index score into an evaluation data set: d (D) 1 ＝(a ij ) m×n Wherein q is the number of questionnaires, n is the index number, n=3, a ij An index score for the j index in the i-th questionnaire;
calculating information entropy of the j index in the i-th questionnaire:
wherein the method comprises the steps of
aj min ＝min{a 1j ,a 2j ,…,a nj }，aj max ＝max{a 1j ,a 2j ,…,a nj }，k＝1/lnq，lnf ij Is f ij Lnq is the natural logarithm of q, when f ij When=0, let f ij lnf ij ＝0；
Calculating the weight of the j index in the i-th questionnaire:
wherein omega (j) is more than or equal to 0 and less than or equal to 1,
calculating the teaching index score of a teacher:
the learning effect index scores are combined into an evaluation data set: d (D) 2 ＝(b i ) p Wherein p is the number of questionnaires, b i Scoring the learning effect index in the ith questionnaire;
calculating a student learning index score:
according to the method, the weights of teaching target index scores, teaching ability index scores and teaching method index scores are calculated through information entropy, then the teacher teaching index score of each questionnaire evaluation result is calculated according to the index weights, and the arithmetic average value is adopted for a plurality of questionnaire investigation evaluation results to obtain the integral teacher teaching index score. The teacher teaching index score is obtained by calculating the combination of the index scores and the questionnaire evaluation results, so that the calculation result is more comprehensive and accurate.
Further, the calculating a final teaching quality evaluation score according to each index score specifically includes:
the teacher teaching index score, the student learning index score, the teaching coordination index score and the teaching order index score are weighted and averaged to obtain a final teaching quality evaluation score which is: score=v 1 Score A1 +v 2 Score A2 +v 3 Score A3 +v 4 Score A4 Wherein v is 1 、v 2 、v 3 And v 4 Respectively the weights of preset teacher teaching indexes, student learning indexes, teaching coordination indexes and teaching order indexes.
The student learning index score is directly obtained by adopting arithmetic average values through a plurality of questionnaire survey evaluation results. The data counted by the questionnaire form is subjective data, but the larger the number of the questionnaires is, the higher the accuracy of the final calculation result is, and in order to obtain a valid result, the number of the questionnaires cannot be 0, so q and p must be greater than 0.
Further, the calculating a final teaching quality evaluation score according to each index score specifically includes:
the teacher teaching index score, the student learning index score, the teaching coordination index score and the teaching order index score are weighted and averaged to obtain a final teaching quality evaluation score which is: score=v 1 Score A1 +v 2 Score A2 +v 3 Score A3 +v 4 Score A4 Wherein v is 1 、v 2 、v 3 And v 4 Respectively the weights of preset teacher teaching indexes, student learning indexes, teaching coordination indexes and teaching order indexes.
According to the invention, subjective data obtained by questionnaire investigation and objective data obtained by artificial intelligent analysis are combined to carry out teaching quality evaluation, and compared with the result obtained by simple questionnaire investigation in the prior art, the result obtained by the simple questionnaire investigation is more comprehensive and objective. The teaching quality evaluation is carried out by three evaluation subjects of teachers, students and artificial intelligence, the teaching effect evaluation indexes and the evaluation principles are considered at multiple angles, the weight proportion of subjective data and objective data can be adjusted according to requirements, and the objectivity and the accuracy of the teaching quality evaluation process and results are ensured.
In a second aspect, an artificial intelligence based classroom teaching assessment system for universities is provided, comprising:
the teaching quality evaluation system construction module is used for constructing a teaching quality evaluation system and comprises the steps of taking a teacher and students as evaluation subjects to construct teacher teaching indexes and student learning indexes, and constructing teaching coordination indexes and teaching order indexes;
the model construction module is used for constructing a teaching coordination index analysis model comprising a text feature extraction module, a voice feature extraction module, a video feature extraction module, a feature fusion module and an emotion recognition module; constructing a teaching order index analysis model comprising a face detection module, a human body detection module and a fusion analysis module;
The model training module is used for manufacturing a training set and a testing set to train and test the teaching coordination index analysis model and the teaching order index analysis model respectively, so as to obtain a trained teaching coordination index analysis model and a trained teaching order index analysis model;
the index data acquisition module is used for acquiring teacher teaching index data and student learning index data, acquiring teaching coordination index data comprising texts, voices and videos and acquiring teaching order index data comprising student image data;
the index score calculation module is used for calculating a teacher teaching index score and a student learning index score according to the teacher teaching index data and the student learning index data;
the index score analysis module is used for analyzing the trained teaching coordination index according to the teaching coordination index data to obtain a teaching coordination index score, and analyzing the trained teaching order index according to the teaching order index data to obtain a teaching order index score;
and the teaching quality evaluation score calculation module is used for calculating a final teaching quality evaluation score according to the index scores.
Compared with the prior art, the invention has the beneficial effects that:
(1) According to the invention, an artificial intelligence technology is applied to teaching quality evaluation, a teaching coordination index analysis model and a teaching order index analysis model are constructed to analyze teaching coordination index data and teaching order index data, objective index scores are obtained as objective data, and subjective data and objective data are combined to carry out teaching quality evaluation, so that the objectivity, scientificity and accuracy of a teaching quality evaluation process and a result can be effectively improved;
(2) According to the invention, emotion recognition is carried out from three dimensions of text features, voice features and video features, and compared with emotion recognition carried out by only one feature, emotion recognition is carried out after feature fusion is carried out on the three features through a teaching coordination index analysis model, so that the obtained result is more accurate;
(3) Aiming at complex situations that students in a class are densely distributed, shielding can influence the recognition effect and the like, the invention provides the teaching order index analysis model for fusing face detection and human body detection, and the model has the characteristics of high robustness and high accuracy, and effectively improves the accuracy of class-to-class rate detection.
Drawings
Fig. 1 is a flow chart of the method of embodiment 1 of the present invention.
Fig. 2 is a diagram of a teaching coordination index analysis model according to embodiment 1 of the present invention.
FIG. 3 is a diagram showing a structure of a teaching order index analysis model according to embodiment 1 of the present invention.
Fig. 4 is a system configuration diagram of embodiment 3 of the present invention.
Detailed Description
The drawings are for illustrative purposes only and are not to be construed as limiting the invention. For better illustration of the following embodiments, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the actual product dimensions; it will be appreciated by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted.
Example 1
As shown in fig. 1, this embodiment provides an artificial intelligence-based classroom teaching evaluation method for universities, including:
s1, constructing a teaching quality evaluation system, wherein a teacher and students are taken as evaluation subjects to construct teacher teaching indexes and student learning indexes, and construct teaching coordination indexes and teaching order indexes;
s2, constructing a teaching coordination index analysis model comprising a text feature extraction module, a voice feature extraction module, a video feature extraction module, a feature fusion module and an emotion recognition module;
s3, constructing a teaching order index analysis model comprising a face detection module, a human body detection module and a fusion analysis module;
S4, a training set and a testing set are manufactured to train and test the teaching coordination index analysis model and the teaching order index analysis model respectively, and a trained teaching coordination index analysis model and a trained teaching order index analysis model are obtained;
s5, collecting teacher teaching index data and student learning index data, collecting teaching coordination index data comprising text data, voice data and video data, and collecting teaching order index data comprising student image data;
s6, calculating a teacher teaching index score and a student learning index score according to the teacher teaching index data and the student learning index data;
s7, analyzing the trained teaching coordination index analysis model according to the teaching coordination index data to obtain a teaching coordination index score, and analyzing the trained teaching order index analysis model according to the teaching order index data to obtain a teaching order index score;
s8, calculating a final teaching quality evaluation score according to each index score;
s9, carrying out teaching quality evaluation according to the final teaching quality evaluation score.
According to the embodiment, an artificial intelligence technology is applied to teaching quality evaluation, artificial intelligence is taken as an evaluation main body, a teaching coordination index analysis model and a teaching order index analysis model are constructed to analyze teaching coordination index data and teaching order index data, objective index scores are obtained to serve as objective data, and subjective data and objective data are combined to perform teaching quality evaluation, so that objectivity, scientificity and accuracy of a teaching quality evaluation process and a teaching order index analysis model can be effectively improved.
The teaching quality evaluation system constructed in the embodiment S1 is provided with a first-level index and a second-level index, where the first-level index includes: teacher teaching index, student learning index, teaching coordination index and teaching order index; wherein the teacher teaching index comprises a second-level index: teaching target indexes, teaching ability indexes and teaching method indexes; the student learning index includes a secondary index: a learning effect index; the teaching coordination indexes comprise secondary indexes: learning emotion indexes and teaching emotion indexes; the teaching order index comprises a second-level index: and (5) a class rate index. The evaluation contents, evaluation subjects, and data attributes of each index are shown in table 1:
table 1 teaching quality evaluation System Table
According to the embodiment, a multi-dimensional teaching quality evaluation system is constructed, the teaching coordination index and the teaching order index which are ignored in the past are introduced into the teaching quality evaluation system, the teaching quality evaluation is carried out through three evaluation subjects of teachers, students and artificial intelligence, and the teaching effect evaluation index and the evaluation principle are considered at multiple angles, so that the teaching quality evaluation dimension is more comprehensive.
The structure diagram of the teaching coordination index analysis model constructed in the step S2 of the embodiment is shown in fig. 2, wherein the text feature extraction module uses a BERT model and a biglu model, the BERT model is an embedding method based on a bidirectional transducer, and the BERT model can capture long-distance dependent features and is the most mainstream and efficient pre-training word embedding model at present. The GRU model is a special cyclic neural network and is simplified by a long and short memory neural network LSTM model, an input gate, a forget gate and an output gate of the LSTM model are combined into an update gate and a reset gate, and two gates of the update gate and the reset gate are used for determining which input information can be finally used as the output of a gating cyclic unit. The bigur model consists of two unidirectional, oppositely directed GRU models, a forward GRU for capturing the context information and an inverse GRU for capturing the context information. In practical use, the BiGRU model of the text feature extraction module is realized by the nn. GRU function of Pytorch, the number of GRU layers of the function is set to be 1, the initial dimension 768 is input, the hidden layer feature dimension is 50, and the text feature vector with the length of 100 is output as a result.
The speech feature extraction module of the teaching coordination index analysis model constructed in step S2 of the embodiment adopts LibROSA speech toolkit, and the embodiment extracts 33 dimension frame-level acoustic features in total, including 1 dimension logarithmic fundamental frequency (log F0), 20 dimension Mel-cepstral coefficient (Mel-Frequency Cepstral Coefficients, MFCCs) and 12 dimension Constant-Qchromatogram (CQT), and further performs speech feature extraction through a biglu model. In actual use, the BiGRU model of the voice feature extraction module is realized by an nn. GRU function of Pytorch, the GRU layer number of the function is set to be 1, the initial dimension 33 is input, the hidden layer feature dimension is 50, and the voice feature vector with the length of 100 is output as a result.
The video feature extraction module of the teaching coordination index analysis model constructed in step S2 of this embodiment extracts face image data using a multitasking convolutional neural network (Multi-task Convolutional Neural Network, MTCNN) face detection algorithm, and then extracts information such as facial markers, facial shape parameters, facial features, head pose, head direction, eye gaze, and the like using a Multi comp openface2.0 toolkit. In this embodiment, 68 facial landmarks, 17 facial action units, a head pose, a head direction, and a set of eye gaze are extracted, and finally frame-level visual features of 709 dimensions are extracted in total, so that video feature extraction is performed through a biglu model. In practical use, the biglu model of the video feature extraction module is implemented by the nn. GRU function of Pytorch, the number of layers of the GRU is set to 1, the initial dimension 709 is input, the hidden layer dimension is 50, and the result is a video feature vector with the length of 100.
The structure diagram of the teaching order index analysis model constructed in step S3 of this embodiment is shown in fig. 3, in which the face detection module adopts the SCRFD algorithm for computing redistribution among different components (backbone, neck and head) of the face detector, and the face detection is performed by using the algorithm, so that the accuracy and efficiency of the model are significantly improved.
In the embodiment, the human body detection module of the teaching order index analysis model constructed in step S3 adopts YOLOv5 algorithm, and the network structure consists of four parts: input part (Input), backbone part (Backbone), neck (ck) and Head (Head). The algorithm has the advantages of high detection speed, high precision, short training time and the like.
In the embodiment, step S4 adopts CH-SIMS to train the teaching coordination index analysis model, and 60 original videos are collected by the CH-SIMS during actual use, and spontaneous expressions, various head postures, shielding and illumination from different movies, television series and variety programs are obtained. Through 60 original videos, 2281 video clips are obtained, the video clips are divided into a training set and a testing set, each video clip is manually marked, marked contents are classified for emotion states, and the method comprises the following steps: negative state (-1), neutral state (0), and positive state (1). And (3) training the model by using a training set, wherein the parameters of the fusion layer and the BiGRU layer are optimized according to the loss function through back propagation and gradient descent during training, so that the loss is minimum until the loss is not reduced any more, and the model is optimal at the moment. In actual use, the optimization function learning rate adam=0.00001 and dropout=0.1 of the model, and the trained teaching coordination index analysis model is obtained after the optimization is completed.
In the embodiment, step S4 collects picture data of classroom students in different time periods, and constructs a data set including scenes such as a class state, an inter-class state and the like. In actual use, 1000 pieces of picture data are collected in total by the data set, and the target detection data set marking tool LabelImg is utilized to mark the picture data so as to divide the types of targets in a classroom scene into: students, tables and stools, establishing classroom student data sets, and dividing the data sets into training sets and test sets. Training the teaching order index analysis model by using a training set through a train training program of YOLOv5, and testing the teaching order index analysis model by using a testing set until a test result meets the standard, thereby obtaining the trained teaching order index analysis model.
Step S5 of this embodiment includes:
q evaluation results reflecting teaching indexes of teachers in the class are obtained through questionnaires, wherein q is more than 0, and each evaluation result comprises a teaching target index score, a teaching ability index score and a teaching method index score;
obtaining p evaluation results reflecting learning indexes of students in a classroom through questionnaires, wherein p is more than 0, and each evaluation result comprises a learning effect index score;
The voice data and the video data of teachers and students in the class are collected by using a sound collection device and a video collection device, and text recognition is carried out on the voice data to obtain text data;
and using a plurality of image acquisition devices with different angles to acquire the image data of students in different time periods in a class.
The present embodiment S6 includes:
s6011, forming an evaluation data set by the teaching target index score, the teaching ability index score and the teaching method index score: d (D) 1 ＝(a ij ) m×n Wherein q is the number of questionnaires, n is the index number, n=3, a ij An index score for the j index in the i-th questionnaire;
s6012, calculating information entropy of the j index in the i-th questionnaire:
wherein->
aj min ＝min{a 1j ,a 2i ,…,a nj }，aj max ＝max{a 1j ,a 2j ,…,a nj }，k＝1/lnq，lnf ij Is f ij Lnq is the natural logarithm of q, when f ij When=0, let f ij lnf ij ＝0；
S6013, calculating the weight of the j index in the i-th questionnaire:
wherein 0.ltoreq.ω (j).ltoreq.1,>
s6014, calculating teacher teaching index scores:
according to the method, a secondary index teaching target index score, a teaching ability index score and a teaching method index score are obtained through questionnaire investigation, weights of the teaching target index score, the teaching ability index score and the teaching method index score are calculated through information entropy, then a teacher teaching index score of each questionnaire evaluation result is calculated according to the secondary index weights, and an arithmetic average value is adopted for a plurality of questionnaire investigation evaluation results to obtain an integral teacher teaching index score.
Step S6 of this embodiment further includes:
s6021, forming the learning effect index score into an evaluation data set: d (D) 2 ＝(b i ) p Wherein p is the number of questionnaires, b i Scoring the learning effect index in the ith questionnaire;
s6022, calculating a student learning index score:
in the embodiment, the secondary index learning effect index score is obtained through questionnaire investigation, and the student learning index score is obtained through arithmetic average of the learning effect index scores of the evaluation results of the questionnaire investigation.
The present embodiment S7 includes:
s7011, a text feature extraction module of a teaching coordination index analysis model is used for encoding text data, and text feature vectors of the encoded text data are extracted;
s7012, processing the voice data by using a voice feature extraction module of the teaching coordination index analysis model, and extracting voice feature vectors of the processed voice data;
s7013, processing the video data by using a video feature extraction module of the teaching coordination index analysis model, and extracting video feature vectors of the processed video data;
s7014, feature fusion is carried out on the text feature vector, the voice feature vector and the video feature vector by using a feature fusion module of the teaching coordination index analysis model;
S7015, carrying out emotion recognition on the fused features by using an emotion recognition module of the teaching coordination index analysis model to obtain a classification result, wherein the classification result comprises a negative state, a neutral state and a positive state;
s7016, obtaining the teaching coordination index score according to the number of the active states of the teachers and the number of the active states of the students identified in a class.
The step S7015 specifically includes:
setting the time of each class as T, and respectively carrying out K emotion recognition on teachers and students in the time T, wherein the number of times that the teachers recognize the positive state is TCount pos The number of times that the student recognizes the positive state is SCount pos The teaching coordination index score is: score A3 ＝(TCount pos +SCount pos )/2K。
In the embodiment, emotion recognition is performed from three dimensions of text features, voice features and video features, and compared with emotion recognition performed by only one of the features, the emotion recognition is performed after feature fusion of the three features, so that the obtained result is more accurate.
The present embodiment S7 further includes:
s7021, a face detection module using a teaching order index analysis model recognizes face detection frame data in an image according to student image data to obtain a student face detection frame set;
s7022, a human body detection module using a teaching order index analysis model recognizes human body detection frame data in an image according to student image data to obtain a student human body detection frame set;
S7023, fusing the student face detection frame set and the student human body detection frame set by using a fusion analysis module of the teaching order index analysis model to obtain a student target detection frame set, and calculating class arrival rates of different time periods in a class according to the student target detection frame set;
s7024, calculating to obtain the teaching order index score according to the class arrival rates of all time periods in a class.
The step S7023 specifically includes:
let the face detection frame set of student be F= { (x) 1 ,y 1 ,w 1 ,h 1 ),…,(x j ,y j ,w j ,h j ),…,(x m ,y m ,w m ,h m ) X, where x j And y j X-axis coordinates and Y-axis coordinates of a center point of a face detection frame of a jth student, w j The width of the face detection frame of the jth student is h j The height of the face detection frame of the jth student is the total number of faces of the student identified by the face detection module;
let the student human body detection frame set be B = { (x) 1 ,y 1 ,w 1 ,h 1 ),…,(x i ,y i ,w i ,h i ),…,(x n ,y n ,w n ,h n ) X, where x i And y i X-axis coordinates and Y-axis coordinates of the center point of the human body detection frame of the ith student, w i Broadband for human body detection frame of ith student, h i N is the height of the human body detection frame of the ith student, and n is the total human body number of the students identified by the human body detection module;
initializing a student target detection frame set as P, and setting P=F, wherein the element number of the set P is m;
Selecting the kth student from the human body detection frame set B, and then B (k) = (x) k ,y k ,w k ,h k )，x k Representing X-axis coordinate, y of human body detection frame of kth student k Represents the Y-axis coordinate, w, of the human body detection frame of the kth student k The width of the human body detection frame of the kth student is represented, h k Representing the height of a human body detection frame of a kth student;
judging whether the human body detection frame of the kth student contains the human face detection frame of any student in the student human face detection frame set F or intersects with the human face detection frame of any student in the student human face detection frame set F, if not, adding B (k) into P;
traversing the human body detection frame set B to obtain a class student target detection frame set P, and setting the element number of the set P as z, namely the current class student number as z;
setting the number of students in a classroom to be N, setting the current time period to be t, and setting the class arrival rate of the time period t to be:
step S7024 specifically includes:
assuming that student image data for M time periods are collected in total, the teaching order index score is:
wherein Q is M Is the lesson arrival rate of time period M.
The core targets of class rate calculation of students in a class are the students and the faces of the students, but because the class scene is bigger, interference objects such as tables and chairs influence human body detection, and the head postures of the students are various during the class, and complete face information is sometimes not observed. Therefore, when the students are positioned, the expected effect of the single face detection or the single body detection on the seriously shielded target is difficult to achieve. Therefore, the embodiment provides the teaching order index analysis model for fusing the face detection and the human body detection aiming at complex situations that the intensive distribution and shielding of classroom students can influence the recognition effect and the like, so that the accuracy of the class-to-class rate detection is effectively improved, and the model has the characteristics of high robustness and high accuracy.
Step S8 of this embodiment includes:
the teacher teaching index score, the student learning index score, the teaching coordination index score and the teaching order index score are weighted and averaged to obtain a final teaching quality evaluation score which is: score=v 1 Score A1 +v 2 Score A2 +v 3 Score A3 +v 4 Score A4 Wherein v is 1 、v 2 、v 3 And v 4 Respectively the weights of preset teacher teaching indexes, student learning indexes, teaching coordination indexes and teaching order indexes.
In the actual calculation process, the weight proportion of subjective data and objective data can be adjusted according to the requirements, so that the objectivity and accuracy of teaching quality evaluation are further ensured.
After the final teaching quality evaluation score is calculated, teaching quality evaluation is carried out according to actual requirements, for example, different teaching quality evaluation result intervals are set according to the teaching quality evaluation score, and a final teaching quality evaluation result is obtained.
According to the embodiment, the artificial intelligence technology is applied to teaching quality evaluation, a multidimensional teaching quality evaluation system is constructed, the subjective index score and the objective index score are combined to carry out teaching quality evaluation, and the objectivity, scientificity and accuracy of a teaching quality evaluation process and a result are effectively improved.
Example 2
In this embodiment, taking the class teaching quality evaluation process of the course "Python programming" as an example, the classroom teaching evaluation method based on artificial intelligence provided in embodiment 1 is applied to specific class teaching quality evaluation.
Constructing a teaching quality evaluation system, which comprises constructing teacher teaching indexes and student learning indexes, and constructing teaching coordination indexes and teaching order indexes by taking a teacher and a student as evaluation subjects; the teacher teaching index data and the student learning index data are collected according to the teaching quality evaluation system, and the specific calculation process is as follows:
in this embodiment, 5 evaluation results reflecting teaching indexes of teachers in a class, namely q=5, are obtained through questionnaires, and are specifically shown in table 2:
| NO | teaching target | Teaching ability | Teaching method |
| 1 | 0.778837 | 0.929208 | 0.929281 |
| 2 | 0.329737 | 0.159161 | 0.267154 |
| 3 | 0.231475 | 0.022563 | 0.555421 |
| 4 | 0.312966 | 0.28622 | 0.036332 |
| 5 | 0.40557 | 0.648136 | 0.174055 |
Table 2 teacher teaching index questionnaire scoring table
According to the formula of the step
Calculating to obtain a teaching target index score, a teaching ability index score and a teaching method index score, wherein the information entropy of the teaching target index score, the teaching ability index score and the teaching method index score is H respectively 1 ＝0.442686979，H 2 ＝0.54388959，H 2 ＝0.289081531。
According to the formula of the step
The weights of the teaching target index score, the teaching ability index score and the teaching method index score are respectively omega (1) = 0.323203317, omega (2) = 0.264512745 and omega (3) = 0.412283938./>
According to the formula of the step
Calculated Score A1 ＝0.57358。
According to the teacher teaching index data and the student learning index data, calculating a teacher teaching index score and a student learning index score, wherein the specific calculation process is as follows:
in this embodiment, 10 evaluation results reflecting learning indexes of students in a class, namely p=10, are obtained through questionnaires, and are specifically shown in table 3:
TABLE 3 student learning index questionnaire scoring sheet
According to the formula of the step
Calculated Score A2 ＝0.379092。
Collecting teaching coordination index data and teaching order index data, inputting the teaching coordination index data into a teaching coordination Score analysis model, and outputting a teaching coordination index Score A3 0.61235 inputting the teaching order index data into the teaching order index Score analysis model, and outputting the teaching order index Score A4 ＝0.68843。
And calculating a final teaching quality evaluation score according to each index score, wherein the specific calculation process is as follows:
in this embodiment, the weight of the teacher teaching index is set to 0.4, the weight of the student learning index is 0.3, the weight of the teaching coordination index is 0.2, the weight of the teaching order index is 0.1, and weighted average is performed to obtain the final teaching quality evaluation Score of score=0.4×0.57358+0.3×0.379092+0.2×0.61235+0.1×
0.68843＝0.534473。
And carrying out teaching quality evaluation according to the final teaching quality evaluation score:
the teaching quality evaluation results can be divided into four standards, namely, the teaching quality evaluation results are poor in 0-0.2, medium in 0.2-0.4, good in 0.4-0.6 and excellent in 0.6-1. According to the final teaching quality evaluation score, the teaching quality evaluation result of the Python programming in the embodiment can be obtained.
Example 3
As shown in fig. 4, this embodiment provides an artificial intelligence-based classroom teaching evaluation system for universities, including:
the teaching quality evaluation system construction module 101 is configured to construct a teaching quality evaluation system, including constructing teacher teaching indexes and student learning indexes, and constructing teaching coordination indexes and teaching order indexes by taking a teacher and a student as evaluation subjects;
the model construction module 102 is used for constructing a teaching coordination index analysis model comprising a text feature extraction module, a voice feature extraction module, a video feature extraction module, a feature fusion module and an emotion recognition module; and constructing a teaching order index analysis model comprising a face detection module, a human body detection module and a fusion analysis module.
The model training module 103 is used for making a training set and a testing set to train and test the teaching coordination index analysis model and the teaching order index analysis model respectively, so as to obtain a trained teaching coordination index analysis model and a trained teaching order index analysis model.
The index data acquisition module 104 is used for acquiring teacher teaching index data and student learning index data, acquiring teaching coordination index data comprising texts, voices and videos and acquiring teaching order index data comprising student image data;
Q evaluation results reflecting teaching indexes of teachers in the class are obtained through questionnaires, wherein q is more than 0, and each evaluation result comprises a teaching target index score, a teaching ability index score and a teaching method index score; and obtaining p evaluation results reflecting the learning index of the students in the class through questionnaires, wherein p is more than 0, and each evaluation result comprises a learning effect index score. The voice data and the video data of teachers and students in the class are collected by using a sound collection device and a video collection device, and text recognition is carried out on the voice data to obtain text data; and using a plurality of image acquisition devices with different angles to acquire the image data of students in different time periods in a class.
An index score calculating module 105 for calculating a teacher teaching index score and a student learning index score from the teacher teaching index data and the student learning index data;
combining the teaching target index score, the teaching ability index score and the teaching method index score into an evaluation data set: d (D) 1 ＝(a ij ) m×n Wherein q is the number of questionnaires, n is the index number, n=3, a ij An index score for the j index in the i-th questionnaire; calculating information entropy of the j index in the i-th questionnaire: 
Wherein->
aj min ＝min{a 1j ,a 2j ,…,a nj }，aj max ＝max{a 1j ,a 2j ,…,a nj }，k＝1/lnq，lnf ij Is f ij Lnq is the natural logarithm of q, when f ij When=0, let f ij lnf ij =0; calculating the weight of the j index in the i-th questionnaire:
wherein 0.ltoreq.ω (j).ltoreq.1,>
calculating the teaching index score of a teacher:
the learning effect index scores are combined into an evaluation data set: d (D) 2 ＝(b i ) p Wherein p is the number of questionnaires, b i Scoring the learning effect index in the ith questionnaire; calculating a student learning index score:
the index score analysis module 106 is used for analyzing the teaching coordination index according to the teaching coordination index data by using the trained teaching coordination index analysis model and obtaining the teaching order index score according to the teaching order index data by using the trained teaching order index analysis model;
encoding the text data by using a text feature extraction module of the teaching coordination index analysis model, and extracting text feature vectors of the encoded text data; processing the voice data by using a voice characteristic extraction module of the teaching coordination index analysis model, and extracting a voice characteristic vector of the processed voice data; processing the video data by using a video feature extraction module of the teaching coordination index analysis model, and extracting the processed video feature vector; feature fusion is carried out on the text feature vector, the voice feature vector and the video feature vector by using a feature fusion module of the teaching coordination index analysis model; carrying out emotion recognition on the fused features by using an emotion recognition module of the teaching coordination index analysis model to obtain a classification result, wherein the classification result comprises a negative state, a neutral state and a positive state; obtaining a Score of the teaching coordination index according to the number of the teacher active states and the number of the student active states identified in a class A3 ；
Use teachingThe face detection module of the learning order index analysis model recognizes face detection frame data in the image according to the student image data to obtain a student face detection frame set; the human body detection module using the teaching order index analysis model recognizes human body detection frame data in the image according to the student image data to obtain a student human body detection frame set; the method comprises the steps that a fusion analysis module of a teaching order index analysis model is used for fusing a student face detection frame set and a student human body detection frame set to obtain a student target detection frame set, and class arrival rates of different time periods in a class are calculated according to the student target detection frame set; calculating to obtain a Score of the teaching order index according to the class arrival rates of all time periods in a class A4 。
A teaching quality evaluation score calculation module 107, configured to calculate a final teaching quality evaluation score according to each index score;
the teacher teaching index score, the student learning index score, the teaching coordination index score and the teaching order index score are weighted and averaged to obtain a final teaching quality evaluation score which is: score=v 1 Score A1 +v 2 Score A2 +v 3 Score A3 +v 4 Score A4 Wherein v is 1 、v 2 、v 3 And v 4 Respectively the weights of preset teacher teaching indexes, student learning indexes, teaching coordination indexes and teaching order indexes.
According to the embodiment, the subjective data obtained by questionnaire investigation and the objective data obtained by model analysis are combined to carry out teaching quality evaluation, and compared with the simple questionnaire investigation in the prior art, the result obtained by the simple questionnaire investigation is more comprehensive and objective. The teaching quality evaluation is carried out by three evaluation subjects of teachers, students and artificial intelligence, the teaching effect evaluation indexes and the evaluation principles are considered at multiple angles, the weight proportion of subjective data and objective data can be adjusted according to requirements, and the objectivity and the accuracy of the teaching quality evaluation process and results are ensured.
It should be understood that the foregoing examples of the present invention are merely illustrative of the present invention and are not intended to limit the present invention to the specific embodiments thereof. Any modification, equivalent replacement, improvement, etc. that comes within the spirit and principle of the claims of the present invention should be included in the protection scope of the claims of the present invention.
Claims (10)
1. The teaching evaluation method for the university class based on the artificial intelligence is characterized by comprising the following steps:
constructing a teaching quality evaluation system, which comprises constructing teacher teaching indexes and student learning indexes, and constructing teaching coordination indexes and teaching order indexes by taking a teacher and a student as evaluation subjects;
Constructing a teaching coordination index analysis model comprising a text feature extraction module, a voice feature extraction module, a video feature extraction module, a feature fusion module and an emotion recognition module;
constructing a teaching order index analysis model comprising a face detection module, a human body detection module and a fusion analysis module;
manufacturing a training set and a testing set to train and test the teaching coordination index analysis model and the teaching order index analysis model respectively, so as to obtain a trained teaching coordination index analysis model and a trained teaching order index analysis model;
collecting teacher teaching index data and student learning index data, collecting teaching coordination index data comprising text data, voice data and video data, and collecting teaching order index data comprising student image data;
calculating a teacher teaching index score and a student learning index score according to the teacher teaching index data and the student learning index data;
analyzing the teaching coordination index according to the teaching coordination index data by using a trained teaching coordination index analysis model, and analyzing the teaching order index according to the teaching order index data by using a trained teaching order index analysis model to obtain a teaching order index score;
Calculating a final teaching quality evaluation score according to the index scores;
and carrying out teaching quality evaluation according to the final teaching quality evaluation score.
2. The teaching evaluation method for universities and colleges according to claim 1, wherein the collecting teaching coordination index data including text data, voice data and video data, and analyzing the teaching coordination index data to obtain a teaching coordination index score by using a trained teaching coordination index analysis model, specifically comprises:
the voice data and the video data of teachers and students in the class are collected by using a sound collection device and a video collection device, and text recognition is carried out on the voice data to obtain text data;
encoding the text data by using a text feature extraction module of the teaching coordination index analysis model, and extracting text feature vectors of the encoded text data;
processing the voice data by using a voice characteristic extraction module of the teaching coordination index analysis model, and extracting a voice characteristic vector of the processed voice data;
processing the video data by using a video feature extraction module of the teaching coordination index analysis model, and extracting the processed video feature vector;
Feature fusion is carried out on the text feature vector, the voice feature vector and the video feature vector by using a feature fusion module of the teaching coordination index analysis model;
carrying out emotion recognition on the fused features by using an emotion recognition module of the teaching coordination index analysis model to obtain a classification result, wherein the classification result comprises a negative state, a neutral state and a positive state;
and obtaining the teaching coordination index score according to the number of the teacher active states and the number of the student active states identified in a class.
3. The method for evaluating classroom teaching in colleges and universities based on artificial intelligence according to claim 2, wherein the method for obtaining the teaching coordination index score according to the number of the teacher positive states and the number of the student positive states identified in a class specifically comprises:
let the time of each class be T, at time TK emotion recognition is respectively carried out on the teacher and the students, wherein the number of times that the teacher recognizes the positive state is TCount pos The number of times that the student recognizes the positive state is SCount pos The teaching coordination index score is: score A3 ＝(TCount pos +SCount pos )/2K。
4. The teaching evaluation method for universities and colleges based on artificial intelligence according to claim 3, wherein the collecting includes face image data and teaching order index data of the human body image data, and the analyzing the teaching order index score according to the teaching order index data using the trained teaching order index analysis model specifically includes:
Using a plurality of image acquisition devices with different angles to acquire student image data in different time periods in a class;
a face detection module using a teaching order index analysis model recognizes face detection frame data in an image according to student image data to obtain a student face detection frame set;
the human body detection module using the teaching order index analysis model recognizes human body detection frame data in the image according to the student image data to obtain a student human body detection frame set;
the method comprises the steps that a fusion analysis module of a teaching order index analysis model is used for fusing a student face detection frame set and a student human body detection frame set to obtain a student target detection frame set, and class arrival rates of different time periods in a class are calculated according to the student target detection frame set;
and calculating according to the class arrival rates of all time periods in a class to obtain the teaching order index score.
5. The method for evaluating classroom teaching in colleges and universities based on artificial intelligence according to claim 4, wherein the fusion module for analyzing the teaching order index fuses the face detection frame set and the body detection frame set of the students to obtain the target set of the student detection frame, and calculates the class arrival rates of different time periods in the class according to the target set of the student detection frame of the classroom, specifically comprising:
Let the face detection frame set of student be F= { (x) 1 ，y 1 ，w 1 ，h 1 )，...，(x j ，y j ，w j ，h j )，...，(x m ，y m ，w m ，h m ) X, where x j And y j X-axis coordinates and Y-axis coordinates of a center point of a face detection frame of a jth student, w j The width of the face detection frame of the jth student is h j The height of the face detection frame of the jth student is the total number of faces of the student identified by the face detection module;
let the student human body detection frame set be B = { (x) 1 ，y 1 ，w 1 ，h 1 )，...，(x i ，y i ，w i ，h i )，...，(x n ，y n ，w n ，h n ) X, where x i And y i X-axis coordinates and Y-axis coordinates of the center point of the human body detection frame of the ith student, w i Broadband for human body detection frame of ith student, h i N is the height of the human body detection frame of the ith student, and n is the total human body number of the students identified by the human body detection module;
initializing a student target detection frame set as P, and setting P=F, wherein the element number of the set P is m;
selecting the kth student from the human body detection frame set B, and then B (k) = (x) k ，y k ，w k ，h k )，x k Representing X-axis coordinate, y of human body detection frame of kth student k Represents the Y-axis coordinate, w, of the human body detection frame of the kth student k The width of the human body detection frame of the kth student is represented, h k Representing the height of a human body detection frame of a kth student;
judging whether the human body detection frame of the kth student contains the human face detection frame of any student in the student human face detection frame set F or intersects with the human face detection frame of any student in the student human face detection frame set F, if not, adding B (k) into P;
Traversing the human body detection frame set B to obtain a class student target detection frame set P, and setting the element number of the set P as z, namely the current class student number as z;
setting the number of students in a classroom to be N, setting the current time period to be t, and setting the class arrival rate of the time period t to be:
6. the method for evaluating classroom teaching in colleges and universities based on artificial intelligence according to claim 5, wherein the method for obtaining the teaching order index score according to the class arrival rate of all time periods in a class specifically comprises the following steps:
assuming that student image data for M time periods are collected in total, the teaching order index score is:
wherein Q is M Is the lesson arrival rate of time period M.
7. The method for evaluating classroom teaching in colleges and universities based on artificial intelligence according to claim 6, wherein the collecting teacher teaching index data and student learning index data specifically comprises:
q evaluation results reflecting teaching indexes of teachers in the class are obtained through questionnaires, wherein q is more than 0, and each evaluation result comprises a teaching target index score, a teaching ability index score and a teaching method index score;
and obtaining p evaluation results reflecting learning indexes of students in a class through questionnaires, wherein p is more than 0, and each evaluation result comprises a learning effect index score.
8. The method for evaluating classroom teaching in colleges and universities based on artificial intelligence according to claim 7, wherein the calculating of the teacher teaching index score and the student learning index score based on the teacher teaching index data and the student learning index data specifically comprises:
combining the teaching target index score, the teaching ability index score and the teaching method index score into an evaluation data set: d (D) 1 ＝(a ij ) m×n Wherein q is the number of questionnaires, n is the index number, n=3, a ij An index score for the j index in the i-th questionnaire;
calculating information entropy of the j index in the i-th questionnaire:
wherein->
aj min ＝min{a 1j ，a 2j ，...，a nj }，aj max ＝max{a 1j ，a 2j ，...，a nj }，k＝1/lnq，lnf ij Is f ij Lnq is the natural logarithm of q, when f ij When=0, let f ij lnf ij ＝0；
Calculating the weight of the j index in the i-th questionnaire:
wherein 0.ltoreq.ω (j).ltoreq.1,>
calculating the teaching index score of a teacher:
the learning effect index scores are combined into an evaluation data set: d (D) 2 ＝(b i ) p Wherein p is the number of questionnaires, b i Scoring the learning effect index in the ith questionnaire;
calculating a student learning index score:
9. the teaching evaluation method for universities and colleges based on artificial intelligence according to claim 8, wherein the calculating the final teaching quality evaluation score according to each index score specifically comprises:
The teacher teaching index score, the student learning index score, the teaching coordination index score and the teaching order index score are weighted and averaged to obtain a final teaching quality evaluation score which is: score=v 1 Score A1 +v 2 Score A2 +v 3 Score A3 +v 4 Score A4 Wherein v is 1 、v 2 、v 3 And v 4 Respectively the weights of preset teacher teaching indexes, student learning indexes, teaching coordination indexes and teaching order indexes.
10. College classroom teaching evaluation system based on artificial intelligence, characterized by comprising:
the teaching quality evaluation system construction module is used for constructing a teaching quality evaluation system and comprises the steps of taking a teacher and students as evaluation subjects to construct teacher teaching indexes and student learning indexes, and constructing teaching coordination indexes and teaching order indexes;
the model construction module is used for constructing a teaching coordination index analysis model comprising a text feature extraction module, a voice feature extraction module, a video feature extraction module, a feature fusion module and an emotion recognition module; constructing a teaching order index analysis model comprising a face detection module, a human body detection module and a fusion module;
the model training module is used for manufacturing a training set and a testing set to train and test the teaching coordination index analysis model and the teaching order index analysis model respectively, so as to obtain a trained teaching coordination index analysis model and a trained teaching order index analysis model;
The index data acquisition module is used for acquiring teacher teaching index data and student learning index data, acquiring teaching coordination index data comprising texts, voices and videos and acquiring teaching order index data comprising student image data;
the index score calculation module is used for calculating a teacher teaching index score and a student learning index score according to the teacher teaching index data and the student learning index data;
the index score analysis module is used for analyzing the trained teaching coordination index according to the teaching coordination index data to obtain a teaching coordination index score, and analyzing the trained teaching order index according to the teaching order index data to obtain a teaching order index score;
and the teaching quality evaluation score calculation module is used for calculating a final teaching quality evaluation score according to the index scores.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310121252.9A CN116362587A (en) | 2023-02-14 | 2023-02-14 | College classroom teaching evaluation method and system based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310121252.9A CN116362587A (en) | 2023-02-14 | 2023-02-14 | College classroom teaching evaluation method and system based on artificial intelligence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116362587A true CN116362587A (en) | 2023-06-30 |
Family
ID=86907473
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310121252.9A Pending CN116362587A (en) | 2023-02-14 | 2023-02-14 | College classroom teaching evaluation method and system based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116362587A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117078094A (en) * | 2023-08-22 | 2023-11-17 | 云启智慧科技有限公司 | Teacher comprehensive ability assessment method based on artificial intelligence |
-
2023
- 2023-02-14 CN CN202310121252.9A patent/CN116362587A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117078094A (en) * | 2023-08-22 | 2023-11-17 | 云启智慧科技有限公司 | Teacher comprehensive ability assessment method based on artificial intelligence |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhu et al. | Visual7w: Grounded question answering in images | |
| WO2020010785A1 (en) | Classroom teaching cognitive load measuring system | |
| Sadjadi et al. | The 2019 NIST Audio-Visual Speaker Recognition Evaluation. | |
| CN108399376A (en) | Student classroom learning interest intelligent analysis method and system | |
| CN107256392A (en) | A kind of comprehensive Emotion identification method of joint image, voice | |
| CN111291613B (en) | Classroom performance evaluation method and system | |
| CN115239527A (en) | Teaching behavior analysis system for teaching characteristic fusion and modeling based on knowledge base | |
| CN110544481A (en) | S-T classification method and device based on voiceprint recognition and equipment terminal | |
| CN116362587A (en) | College classroom teaching evaluation method and system based on artificial intelligence | |
| CN110046271A (en) | A kind of remote sensing images based on vocal guidance describe method | |
| CN116304973A (en) | Classroom teaching emotion recognition method and system based on multi-mode fusion | |
| CN116050892A (en) | Intelligent education evaluation supervision method based on artificial intelligence | |
| Petrova et al. | Group-level emotion recognition using a unimodal privacy-safe non-individual approach | |
| CN110956142A (en) | Intelligent interactive training system | |
| CN111611854A (en) | Classroom condition evaluation method based on pattern recognition | |
| CN113283334B (en) | Classroom concentration analysis method, device and storage medium | |
| CN117078094A (en) | Teacher comprehensive ability assessment method based on artificial intelligence | |
| Chen et al. | Intelligent teaching evaluation system integrating facial expression and behavior recognition in teaching video | |
| CN116244474A (en) | Learner learning state acquisition method based on multi-mode emotion feature fusion | |
| Zou et al. | Detecting practical speech emotion in a cognitive task | |
| CN115905977A (en) | System and method for monitoring negative emotion in family sibling interaction process | |
| Chen et al. | Intelligent Recognition of Physical Education Teachers' Behaviors Using Kinect Sensors and Machine Learning. | |
| CN111950472A (en) | Teacher grinding evaluation method and system | |
| CN112818741A (en) | Behavior etiquette dimension evaluation method and device for intelligent interview | |
| CN111081095A (en) | Video and audio teaching platform, analysis subsystem and method, recommendation subsystem and method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |