Embedding Affect Awareness into Online Learning Environment using Deep Neural Network

E-learning has developed speedily in past years, and has become a trend of learning thanks to luxury of using it at anytime, anywhere. Engineering educators consider Bloom's taxonomy to make learning objectives. According to Bloom's taxonomy teaching objectives can be characterized into 3 domains: cognitive, affective and psychomotor domain. While assessing the students in E-learning, most instructing courses give weight on information impartation and neglect students' learning attitudes and also the affective domain because of the separation. An emotion plays a very important role in cognitive process of human being, managing learning activities and it reflects on the studies. The present E-learning frameworks regularly lack many features of profound affection, and neglect to give appropriate passionate association. The methodology presented in this paper enables real time detection of individual's affective states using non wearable sensors. This study explores however feeling evolves throughout the educational method. Our system recognizes seven different emotions such as Angry, Disgust, Fear, Happy, Sad, Surprise and Contempt and classifies them into positive, negative and neutral emotions using Convolution Neural Network (CNN). Also interaction is carried out between system and learner to solve the learner's doubts related to understanding of the topic in video.