Machine Learning based Drowsiness Detection in Classrooms

The Major objective of this proposed work is to detect the mood swings of face and drowsiness especially in schools and graduate levels. This work is supported by the machine learning and various studies on existing methods of drowsiness detection suggested the impact of artificial intelligence in i...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:2022 International Conference on Edge Computing and Applications (ICECAA) s. 1186 - 1191
Hlavní autori: R, Dinesh Kumar J, N, Harish, K, Priyadharsini, S, Gowtham, Gokulraj, Gokulraj
Médium: Konferenčný príspevok..
Jazyk:English
Vydavateľské údaje: IEEE 13.10.2022
Predmet:
Computational modeling
Drowsiness Detection
Face recognition
Haar Cascade algorithm
Machine Learning
Machine learning algorithms
Mood
OpenCV
Radio frequency
Support vector machines
XML
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:The Major objective of this proposed work is to detect the mood swings of face and drowsiness especially in schools and graduate levels. This work is supported by the machine learning and various studies on existing methods of drowsiness detection suggested the impact of artificial intelligence in identifying the face mood swings. Thisproposed system follows the machine learning models and it shows a better performance. The system has been implemented in classrooms. The base of thispaperwork involves OpenCV library of python which acts as a great tool for image processing, computer vision, and object tracking. The Haar cascade algorithm is used for face detection which recognizes faces of students in real-time video and follows the stages of Haar transformations. The Haar cascade uses the eyeball movement as a parameter is distinguish the various mood swings of the students. The main intention of thisproposed system is to reduce the drowsiness level among the students in classroom environment by providing alert feedback at centralised manner. The proposed model has 85.3% and 95.3% accuracies for detecting the mood swings and eye recognition. The models are developed and tested on SVM, NN, KNN and RF where the NN has the maximum accuracy of 84% and this work shows the overall improvement of 15% in all the primary aspects.
DOI:10.1109/ICECAA55415.2022.9936550