Eye State Detection Based Students’ Attention Assessment In Online Learning

In recent years,online learning,especially Massive Open Online Course(MOOC),has developed rapidly.Online learning eliminates certain obstacles which are unavoidable in traditional learning,such as differences in people’s age,physical conditions,and individual acceptance,so that any learners can learn anything at anywhere and anytime.In addition,the development of ICTs(information and communications technologies)and multimedia-based technologies create opportunities and conditions for monitoring,analyzing and predicting trends and patterns in student learning behavior.This type of analysis can help teachers design a new and effective way to teach and deliver teaching content in a better way.The separation of teachers and students are separated by cyberspace in online learning,causes the lock of emotional communication between them.It is easy for students to produce negative emotions such as disgusting students,and lead students to lose concentration or give up learning.Therefore,how to identify the learner’s attention state in online learning,and get effective feedback on the learner’s cognition and emotion,plays an important role in making more targeted and attractive learning contents.The aim of this study aims is to propose a student visual attention state assessment algorithm by analyzing the video sources and extracting visual attention features.In order to assess the students’ attention status,this paper first solves the problems of student’s face detection,facial points localization,and eye status detection.The main work of this paper includes:(1)Localization of student’s eyes: In order to localize the student’s eyes,firstly we used Adaboost and supervised descent model(SDM)to detect human face and extract68 feature points of face.Through a lot of experiments,we found that the Adaboost and SDM can achieve a good performance for the localization of student’s eyes.These results laid the foundation for the eye state detection and attention state accessment.(2)Student eye state detection: An eye state recognition algorithm based on Gabor and support vector machine(SVM)is proposed.The Gabor filter extracts the detail features representing the eye and then classifies the features using SVM.Through experimental comparison,we found that Gabor and SVM are superior to existing algorithms.(3)Student’s attention status assessment: We combine the localization of eyes,human eye status detection.The student’s attention status is scored using the PERCLOS criteria.The feasibility of the algorithm is proved by simulation experiments.