Open AccessDeep feature based efficient regularised ensemble for engagement recognition
Author(s) -
Park Y.m.,
Lee G.M.,
Yang H.S.
Publication year - 2019
Publication title -
electronics letters
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.375
H-Index - 146
eISSN - 1350-911X
pISSN - 0013-5194
DOI - 10.1049/el.2019.2783
Subject(s) - bridge (graph theory) , computer science , artificial intelligence , feature (linguistics) , deep learning , set (abstract data type) , machine learning , artificial neural network , pattern recognition (psychology) , medicine , philosophy , linguistics , programming language
Over the years, open education in online environments, such as Massive Online Open Courses, has grown rapidly. While the trend is expected to bridge the educational gap among students, the new environment has also created new challenges such as the lack of feedback and difficulties in interaction. The authors propose an automated engagement recognition system to alleviate this problem, driven by the recent developments in computer vision and artificial neural networks. The authors' proposed system extracts deep features from a facial image and employs a combination of multiple regularised shallow networks to recognise engagement. They verified the system in a public data set. The proposed system has faster learning speed and better accuracy than single deep network based approaches do.