Abnormal Behavior Determination Model of Multimedia Classroom Students Based on Multi-task Deep Learning

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216305%3A26220%2F23%3APU148924" target="_blank" >RIV/00216305:26220/23:PU148924 - isvavai.cz</a>

Výsledek na webu

<a href="https://link.springer.com/article/10.1007/s11036-023-02187-7" target="_blank" >https://link.springer.com/article/10.1007/s11036-023-02187-7</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1007/s11036-023-02187-7" target="_blank" >10.1007/s11036-023-02187-7</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Abnormal Behavior Determination Model of Multimedia Classroom Students Based on Multi-task Deep Learning

Popis výsledku v původním jazyce

The abnormal behavior of students in the multimedia classroom is not significant, which leads to the difficulty in determining abnormal behavior. Therefore, the abnormal behavior determination model of multimedia classroom students based on multi-task deep learning is constructed. The eigenimage filtering algorithm is used to denoise the captured multimedia classroom student images. The multimedia classroom student images are denoised using an adaptive histogram equalization algorithm to enhance the denoised multimedia classroom student images. The multimedia classroom student images are segmented using the Renyi entropy method, and the student behavioral characteristics are determined based on the image segmentation results. Student behavioral characteristics are determined based on image segmentation results. A multi-task deep learning model is built based on convolutional neural networks. The model mainly uses convolutional neural networks and students' behavioral features to classify students' abnormal behaviors in multimedia classrooms, achieve the determination of abnormal behaviors of multimedia classroom students, and obtain relevant determination results. The experimental results show that the model can effectively determine the abnormal behaviors of students in multimedia classrooms, such as looking to the right and looking left, playing with mobile phones, etc. The accuracy of the determination of abnormal behavior is higher than 98%, and the practical application is good.

Název v anglickém jazyce

Abnormal Behavior Determination Model of Multimedia Classroom Students Based on Multi-task Deep Learning

Popis výsledku anglicky

The abnormal behavior of students in the multimedia classroom is not significant, which leads to the difficulty in determining abnormal behavior. Therefore, the abnormal behavior determination model of multimedia classroom students based on multi-task deep learning is constructed. The eigenimage filtering algorithm is used to denoise the captured multimedia classroom student images. The multimedia classroom student images are denoised using an adaptive histogram equalization algorithm to enhance the denoised multimedia classroom student images. The multimedia classroom student images are segmented using the Renyi entropy method, and the student behavioral characteristics are determined based on the image segmentation results. Student behavioral characteristics are determined based on image segmentation results. A multi-task deep learning model is built based on convolutional neural networks. The model mainly uses convolutional neural networks and students' behavioral features to classify students' abnormal behaviors in multimedia classrooms, achieve the determination of abnormal behaviors of multimedia classroom students, and obtain relevant determination results. The experimental results show that the model can effectively determine the abnormal behaviors of students in multimedia classrooms, such as looking to the right and looking left, playing with mobile phones, etc. The accuracy of the determination of abnormal behavior is higher than 98%, and the practical application is good.

Druh

J_imp - Článek v periodiku v databázi Web of Science

CEP obor

—

OECD FORD obor

20200 - Electrical engineering, Electronic engineering, Information engineering

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach

Rok uplatnění

2023

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

MOBILE NETWORKS & APPLICATIONS

ISSN

1383-469X

e-ISSN

1572-8153

Svazek periodika

neuvedeno

Číslo periodika v rámci svazku

neuvedeno

Stát vydavatele periodika

NL - Nizozemsko

Počet stran výsledku

14

Strana od-do

„“-„“

Kód UT WoS článku

001050927300003

EID výsledku v databázi Scopus

2-s2.0-85168327067