Abstract:Classroom environment is a competent platform for the students to learn and improve their understanding of the subject. An instructor’s primary responsibility lies in managing the students in a way they feel interested and focused during the class. With the aid of automated systems based on artificial intelligence, an instructor can get feedback on the students’ attention span in the class by monitoring their emotions using learning algorithms that can prove to be effective to improve the teaching style of the instructor that can in turn have positive effects on the class. In this paper, we propose an LSTM recurrent neural network trained on an emotional corpus database to extract the speech features and convolutional neural networks trained on the FER2013 facial emotion recognition database were used to predict the speech and facial emotions of the students respectively, in real-time. The live video and audio sequence of the students captured is fed to the learned model to classify the emotions individually. Once the emotions such as anger, sadness, happiness, surprise, fear, disgust and neutral were identified, a decision-making mechanism was used to analyze the predicted emotions and choose the overall group emotion by virtue of the highest peak value achieved. This research approach has the potential to be deployed in video conferences, online classes etc. This implementation proposal should effectively improve the classification accuracy and the relatability of the detected student emotions and facilitate in the design of sophisticated automated learning systems that can be a valuable tool in evaluating both the students and the instructors. The adapted research methodologies and their results are discussed and found to perform suggestively better than the other research works used in the comparison

Emotion Recognition of College Students' Online Learning Engagement Based on Deep Learning

Emotion Detection in Online Social Networks: A Multilabel Learning Approach

Student Sentiment Analysis and Classroom Feedback Prediction Using Deep Learning

Machine Learning-Based Student Emotion Recognition for Business English Class

CNN for Facial Emotion Recognition in Online Learning Platforms to Identify Learner Engagement

Emotion recognition and achievement prediction for foreign language learners under the background of network teaching

Enhancing Student Engagement in Online Learning through Facial Expression Analysis and Complex Emotion Recognition using Deep Learning

Multimodal Engagement Recognition From Image Traits Using Deep Learning Techniques

Study on the Distance Learners’ Academic Emotions Using Online Learning Behavior Data

Facial emotion recognition based real-time learner engagement detection system in online learning context using deep learning models

Multimodal Emotion Recognition and State Analysis of Classroom Video and Audio Based on Deep Neural Network

Extraction and analysis of factors influencing college students’ mental health based on deep learning model

Classifying Emotions and Engagement in Online Learning Based on a Single Facial Expression Recognition Neural Network

Exploring AI Facial Recognition for Real-time Emotion Detection: Assessing Student Engagement in Online Learning Environments

A Deep Learning-based Emotion Recognition System for Interactive E-Learning

Emotion Analysis Method of Teaching Evaluation Texts Based on Deep Learning in Big Data Environment

Advances and Application of Facial Expression and Learning Emotion Recognition in Classroom

Emotion recognition of social media users based on deep learning

A Multimodal Intelligent Emotion Perception Framework by Data-driven and Knowledge-guided

Classroom Student Emotions Classification from Facial Expressions and Speech Signals using Deep Learning

Deep Learning Based Emotion Recognition and Visualization of Figural Representation