Research Area:  Machine Learning

Abstract:

In every learning setting, in classrooms or online, a student-s emotions throughout course involvement play a critical role. It employs disturbing, excite, and eye and head movement patterns to infer important information about a student-s mood in an e-learning environment. Researchers from numerous disciplines have been focusing on emotion detection technologies to better understand user engagement, efficacy, and utility of systems that have been established or are being deployed. The goal of this study is to see if students facial expressions can be used by lecturers to understand students comprehension levels in a virtual classroom, as well as to determine the influence of facial expressions during lectures and the degree of comprehension displayed by these emotions. The objective is to determine which facial physical behaviours are associated with emotional states and then to determine how these emotional states are related to student understanding. The major purpose of this work is to plan and develop a new deep learning-oriented facial expression recognition (FER) of online learners from real-time videos. For the frames of online learners from real-time videos, the Viola–Jones Face detection algorithm is employed for face detection. Further, the pattern extraction is performed by the optimized local directional Texture Pattern (LDTP) using the hybrid Coyote Optimization Algorithm (COA), and Deer Hunting Optimization Algorithm (DHOA) referred as Coyote–Deer Hunting Optimization (C-DHO). These pattern images are inputted to the convolutional neural network (CNN) for deep feature extraction. Furthermore, the heuristically modified recurrent neural network (HM-RNN) using the same C-DHO is used for the expression recognition. The experimental research reveals that the suggested method aids in the identification of emotions as well as the classification of student participation and interest in the topic, all of which are displayed as feedback to the teacher in terms of improving the learner experience.

Keywords:  
Facial Expression Recognition
Deep Learning Model
real-time videos
local directional Texture Pattern (LDTP)
hybrid Coyote Optimization Algorithm (COA)

Author(s) Name:  M. Jagadeesh & B. Baranidharan

Publisher name:  Springer

DOI:  10.1007/s00530-022-00957-z

Volume Information:  

Paper Link:   https://link.springer.com/article/10.1007/s00530-022-00957-z