Enhancement of Student Attentiveness Using Deep Learning Techniques

• Published in: International journal of e-collaboration 2022-01, Vol.18 (3), p.1-14
• Main Authors: Aruna, S, Kuchibhotla, Swarna
• Format: Article
• Language: English
• Subjects: Accuracy; Analysis; Artificial neural networks; Collaboration; Datasets; Deep learning; Education; Emotion recognition; Emotions; Equipment and supplies; Eye movements; Image enhancement; Image processing; Image quality; Machine learning; Methods; Neural networks; Technology application; Tracking systems
• ISSN: 1548-3673; 1548-3681
• DOI: 10.4018/IJeC.307129

The cognitive state of mind helps us to perceive various kinds of information which can aid us in inferring various insights. Emotion recognition has been a prominent field of study which helps us to get insights into the cognitive state of mind. There have been various prior works done in the field of emotion recognition which has their applications in the fields of education, marketing, analysis, etc. Having access to such information allows us to draw insights. Though there have been various models made, there hasn't been much focus on improving the features of an image that define an emotion. The paper which we propose is to use image processing techniques on the images which would enhance the quality of the images and make them run under a convolutional neural network (CNN) along with eye tracking system to track the gaze of a student in order to identify the attentivity of a student. The dataset being used for this experiment is the FEC dataset which contains a set of 35000 images of 48 x 48 size. Experimentally determined results have resulted in achieving an accuracy of 94%.