43
3

Multimodal Group Activity Dataset for Classroom Engagement Level Prediction

Abstract

We collected a new dataset that includes approximately eight hours of audiovisual recordings of a group of students and their self-evaluation scores for classroom engagement. The dataset and data analysis scripts are available on our open-source repository. We developed baseline face-based and group-activity-based image and video recognition models. Our image models yield 45-85% test accuracy with face-area inputs on person-based classification task. Our video models achieved up to 71% test accuracy on group-level prediction using group activity video inputs. In this technical report, we shared the details of our end-to-end human-centered engagement analysis pipeline from data collection to model development.

View on arXiv
Comments on this paper

We use cookies and other tracking technologies to improve your browsing experience on our website, to show you personalized content and targeted ads, to analyze our website traffic, and to understand where our visitors are coming from. See our policy.