Font Size: a A A

Interest Evaluation Of Learners Based On Multidimensional Feature Fusion

Posted on:2020-05-29Degree:MasterType:Thesis
Country:ChinaCandidate:T ZhangFull Text:PDF
GTID:2428330575974269Subject:Information and Communication Engineering
Abstract/Summary:PDF Full Text Request
With the continuous development of the times,the demands and expectations of society for education are constantly improving.The educational model and evaluation model which used to focus on knowledge and memory are no longer suitable for the needs of education in the new era.It has become an urgent problem for educational authorities,schools at all levels and teachers to find out their potential direction and develop individualized in their suitable fields with a high degree of interest.The evaluation of students' interest in the learning process can help students explore the potential direction of personal learning,the true direction of interest and better understand their own learning needs.For teachers,they can better understand the learning status of students,conduct targeted teaching guidance for students,improve teaching efficiency,and realize teaching in accordance with their aptitude.Based on the theoretical research on academic emotions,body posture and students'academic interest,this paper adopts four dimensions of facial emotion,head posture,body sitting posture and eye posture;,and uses deep learning and machine learning algorithm to realize the evaluation and analysis of students'interest in learning.Specifically include:1.For the teaching activity video,the key frame selection is carried out;using the openpose platform,the key points of the face and the key points of the head and trunk joints are obtained,which provides basic data support for the learning interest feature extraction.2.Based on educational theories such as academic emotions and learning state analysis,several dimensional features and extraction methods related to learning interest,such as learning emotions,head posture and body sitting posture,are established.Aiming at academic sentiment,a calculation method of learning emotion characteristics based on pleasure intensity statistics is proposed.3.Principal component analysis method was used to reduce and integrate the multi-dimensional features related to learning interest,and the feature model of learning interest evaluation was established.4.Based on the feature model of learning interest degree evaluation,a learning interest degree evaluation method based on SVM is proposed.
Keywords/Search Tags:Learning interest, Academic mood, Sitting posture, Head posture, SVM
PDF Full Text Request
Related items