Research And Implementation Of Students’ Learning State Monitoring Method Based On Posture Analysis

With the continuous updating of teaching mode in our country,the pressure of teachers in teaching activities is also increasing.In order to maximize the learning ability of each student in the group,it is necessary for the teacher to keep the attention to each student.However,in traditional classroom teaching,the teacher can not monitor the learning state of each student,this leads to some students are not paid attention by teachers,resulting in the polarization of the better and the worse,which is an urgent problem to be solved in traditional teaching.The main content of this paper is to realize the classification of students’ learning state in classroom learning through the research on the recognition of students’ learning posture,and to design the classification system of students’ learning state.The specific content of the study is as follows:1.According to the characteristics of three learning states of concentration,generality and non-concentration in classroom learning,six characteristic actions of identifying learning states are put forward.And,the learning state feature space model is designed.2.This paper analyzes the learning posture characteristics of students in different learning states,and puts forward a learning state recognition method based on learning concentration score.3.A learning state classification recognition algorithm based on convolutional neural network is designed,and the network structure is designed,the network parameters are debugged and the network is trained.And The influence of each parameter setting on the result is tested,and the optimal network structure is found.The average recognition accuracy reaches 81.5%,and the recognition accuracy with an error of ±0.14 is obtained.4.A learning state classification recognition algorithm based on support vector machine is designed,and six recognition methods of characteristic actions are described.It is proposed to classify learning state by using the output probabilities of three support vector machines.Finally,through the learning state classification experiment based on support vector machine,the average recognition accuracy of learning state reaches 78.6%.5.In this paper,we combine the learning state classification recognition algorithm based on the convolutional neural network with the learning state classification recognition algorithm based on support vector machine,the results show that the accuracy of fusion algorithm is better than that of single classification algorithm,and the average accuracy reaches 83.3%.6.The learning state monitoring and classification system based on students’ learning posture in class is completed.