Prediction of Students' Psychological States Based on Facial

Expression Recognition Under the Influence of COVID-19

Huiling Liu

, Peiyu Liu, Ran Lu

and Shuqin Li

School of information science and Engineering, Shandong Normal University, Jinan, China

Keywords: COVID-19, Facial Expression Recognition, Psychological Prediction, Higher Education, Online Learning.

Abstract: Affected by COVID-19, people have to face long-term and wide-ranging social group isolation. College

students, as one of the important groups in society, have converted to online learning systems, which may

lead to various psychological and emotional problems. Their mental health problems should be paid special

attention to. In this article, we utilize facial expression recognition technology to predict students'

psychological problems during online learning, with the international Self-Rating Anxiety Scale (SAS) as an

auxiliary prediction tool. Experimental results show that it is feasible to use facial expression recognition

technology to predict students' psychological problems. In the end, we gave intervention measures to students

with different anxiety levels under the influence of the epidemic from a scientific perspective to help students

adjust their mentality in time and return to normal study and life.

1 INTRODUCTION

At the beginning of 2020, coronavirus pandemic

struck world. The World Health Organization

announced that the COVID-19 disease caused by

SARScov-2 coronavirus is a mass disease, which

poses a broad public health threat all over the world

(Panskyi et al., 2021). In January 2020, the Chinese

government entered a state of emergency: areas with

high risk of the epidemic implemented a "blockade"

and areas with low and medium risk of the epidemic,

restrictions on social and economic life were

imposed, including prohibition of leaving the place of

residence, isolation at home, restrictions on engaging

in economic or professional activities. Of course, this

also meant the closure of schools and other

educational institutions. In order to ensure the health

and safety of teaching and students, all kinds of

schools at all levels, especially higher education

schools, made use of the advantages provided by

information technology to carry out distance

education, which made students' learning mode

change to online learning (Alzahrani et al., 2021).

A few months later, the epidemic situation has

been significantly controlled, but people still have to

face prolonged isolation. Students, as one of the

https://orcid.org/0000-0002-0254-4840

important groups in society, because of the unsealing

of the school, are facing emergency remote learning

(ERL) (Tulaskar et al., 2021) and waiting for school

start. Due to the hasty transition to distance learning,

students lack experience and are not familiar with

new online resources, which may lead to various

physical, psychological and emotional problems.

Their mental health problems should be paid special

attention to (Fu et al., 2021). In this paper, our work

is to accurately identify the mental health problems of

students, timely help students adjust their mentality

and restore their normal study and life state during the

epidemic period.

Mood and facial muscle fluctuations are called

facial expressions. It provides us with clues about a

person's state and enables us to talk to each other

according to their emotions (Luqin et al., 2019).

Facial expression is a spontaneous psychological

state, without conscious efforts to make physiological

changes in facial muscles. Facial expression is a very

important means of nonverbal communication, which

can express people's inner emotions and emotions

(Calbi et al., 2021). For students, they may frown and

close their lips when they feel uncomfortable. Timely

identification of students' facial expression is helpful

to understand students' psychological state in time.

Liu, H., Liu, P., Lu, R. and Li, S.

Prediction of Students’ Psychological States Based on Facial Expression Recognition Under the Inﬂuence of COVID-19.

DOI: 10.5220/0011909600003613

In Proceedings of the 2nd International Conference on New Media Development and Modernized Education (NMDME 2022), pages 235-243

ISBN: 978-989-758-630-9

 2023 by SCITEPRESS – Science and Technology Publications, Lda. Under CC license (CC BY-NC-ND 4.0)

235

Emotions are divided into seven categories:

happiness, anger, fear, sadness, disgust, surprise and

nature (see Figure 1) (Kumar et al., 2017). The main

purpose of this paper is to use a facial expression

recognition technology based on convolutional neural

network to predict students' psychological problems

by recognizing students' facial expressions. In the

experiment, we recruited 100 college student

volunteers to collect facial expressions through the

network teaching platform, and input the collected

pictures into the pre training model for facial

expression recognition. After the experiment, the

international Self-Rating Anxiety Scale (SAS) was

used to test the psychology of volunteers. At the same

time, volunteers were asked to report whether they

were distracted and whether their emotions

fluctuated, which was taken as the goal of the

prediction model. Through comprehensive analysis,

53% of the volunteers had different degrees of

anxiety. It was learned from the interview that this

was directly related to the isolation of home and

school caused by the epidemic. Finally, we give

corresponding suggestions for students'

psychological problems.

Figure 1. (1) is Angry face, (2) is disgusted face, (3) is

fearful face, (4) is happy face, (5) is sad face, (6) is surprised

face, (7) is natural face

The rest of this paper is structured as follows. The

application scenarios and the latest development of

facial expression recognition technology are

discussed in Sect. 2. In Sect. 3, the facial expression

recognition experiment and model algorithm based

on convolutional neural network are described. In

Sect. 4, the experimental analysis and SAS are

discussed to prove the efficiency of students'

psychological problems prediction based on facial

expression recognition technology. The causes and

corresponding suggestions of students' psychological

problems brought by COVID-19 are discussed in

Sect. 5. Finally, the conclusion of the paper is

presented in Sect. 6.

2 RELATED WORK

Charles Darwin was one of the first scientists to

realize that facial expression is one of the most

powerful and direct ways for human beings to

exchange emotions, intentions and opinions (Bartlett

et al., 2003). Facial expressions not only provide

emotional state information, but also show cognitive

state information such as interest, boredom,

confusion and stress. John Gottman, an American

psychologist, believes that from the perspective of

individual psychology, human facial expressions are

uncontrollable and subconscious behavior. For

example, when you see terrible things, you will feel

afraid; when you see delicious food, you will drool

subconsciously.

Nowadays, facial expression recognition

technology has been applied to different fields,

including human-computer interaction, security,

robot manufacturing, medical treatment, games and

automobile manufacturing (Căleanu et al., 2013).

In terms of security, such as the lie detection work

of criminals, some criminals are good at camouflage

when they are interrogated. But some criminal

psychologists believe that after repeating the same

problems, criminals will become tired and reveal their

flaws. Misjudgment may be caused by different

evaluation criteria and expression omission if police

only rely on human observation. So it is necessary to

use a computer with expression recognition ability to

assist in judgment (Li et al., 2021). In the field of

automobile, facial expression recognition mainly

carries out for drivers to judge whether they are tired

or not. The drivers’ state behind the expression is

analyzed by using the camera to capture the driver's

facial expression features in real time (Chanchal et

al., 2016). Once it is found that the drivers may be

tired driving (eye closure timeout, drooping

eyebrows, passive eye opening, etc.), the driving

system will start a series of intervention means (Jabon

et al., 2010), such as voice broadcast prompt, playing

loud music to refresh the driver, etc. Zhao et al.

proved the accuracy of driver fatigue prediction by

using face recognition technology (Zhao et al., 2018).

In robot manufacturing and medical treatment, such

as in clinical treatment and service industry, the

combination of facial expression recognition can

make the robot better understand people's

psychological activities, so as to provide people with

more accurate and thoughtful services (Ilyas et al.,

2021). In terms of games, In the field of education,

face recognition technology is widely used at present.

For example, face recognition technology can be used

to confirm the identity of candidates in various major

examinations in universities to prevent substitute

exams. However, the research on students' facial

expression of teachers' teaching and students' learning

is relatively few. Emotion has a positive effect on

motivation. Students' classroom emotion is closely

related to learning motivation and is one of the

important factors affecting learning effect.

NMDME 2022 - The International Conference on New Media Development and Modernized Education

236

Students' facial expression is the most effective

way to express emotion. Facial expression is the main

source of information second only to speech when

determining a person's inner feelings. Teachers can

interpret students' emotions by paying attention to the

changes of students' facial expressions to adjust their

emotional state in time. But affected by the number

of classes, teachers' ability and experience, the effect

is poor (Mohiadeen et al., 2021). With the wide

application of intelligent technology in education,

schools have built smart classrooms, which can

capture facial features in real time to recognize their

facial expressions, so as to analyze the learning state

of students.

Under the influence of COVID-19, depression,

panic and anxiety have been a serious problem for

people due to lack of knowledge. Especially for

college students, online studying has great impact on

their learning habits. Many students even have some

psychological pressure, anxiety and irritability due to

academic pressure and employment pressure. With

the global epidemic continuing, it is very important to

pay attention to the mental health education of college

students during the epidemic.

3 METHODOLOGY

We use algorithm model provided by GitHub user ID

"WuJie1010". This algorithm uses convolutional

neural network to integrate facial expression feature

extraction and expression classification into an end-

to-end network. VGG19 consists of 16 convolution

layers and the last 3 full connection layers. The

algorithm removes multiple full connection layers in

the traditional VGG19 and directly divides it into

seven categories after a full connection layer. The

specific code can be found in GitHub

https://github.com/WuJie1010/Facial-Expression-

Recognition.Pytorch.

3.1 Dataset

We continued to use the CK+ dataset. CK+ dataset is

an extension of CK dataset and a standard database in

facial expression recognition. Many studies will use

this data for testing. It contains 327 tagged facial

videos. We extracted fragments from the CK +

dataset, which contains a total of 981 facial

expressions. In addition, we also performed data

enhancement operations. On the CK+ dataset, we

artificially do some image transformations such as

flipping and rotation, as shown in Figure 2, so as to

prevent data over fitting, expand the amount of data

in the database and make the trained network more

robust.

Figure 2. Data enhancement operation

3.2 Model Training

In this process, we divide the image data set into

training set and test set. The eigenvalues are

calculated in the training set, and the features are

divided into 7 categories. The pre-training model is

generated through 300 rounds of learning and

training, and verified by the test set. Through the

experimental evaluation, good results are obtained,

and the accuracy has a certain reference value for the

future research of computer-based emotion

recognition system model. The accuracy of facial

expression recognition of this model is as high as

93.57%. As shown in Table 1.

Table 1. Model accuracy

Model

Accuracy

Training Testing

VGG 98.7% 93.57%

Then, we input the test pictures into the best

model trained under CK + data set, get the

probabilities of different expressions, and then

visualize the probabilities of pictures in different

classifications and the results predicted by the model.

As shown in Figure 3.

Prediction of Students’ Psychological States Based on Facial Expression Recognition Under the Inﬂuence of COVID-19

237

Figure 3. Visualization results of facial expression recognition

4 EXPERIMENT

We recruited 100 college student volunteers (20< age

<25) to collect their facial expressions through the

network teaching platform and input the collected

pictures into the pre-training model for facial

expression recognition. Volunteers were asked to

report whether they were distracted and whether their

emotions fluctuated, which was the goal of the

prediction model. In order to better evaluate the

psychological state of volunteers, the international

Self-rating Anxiety Scale was used to test the

psychological anxiety of volunteers after the

experiment.

4.1 Acquisition of Facial Expressions

We invited teachers' volunteers and students'

volunteers to use the online learning platform for

online teaching and studying. During the teaching

process, we continuously intercepted the learning

status of students at a certain interval. Some of them

are shown in Figure 4. We cut and flip the obtained

student pictures, and finally obtained 100 student

facial expression pictures.

NMDME 2022 - The International Conference on New Media Development and Modernized Education

238

Figure 4. Acquisition of students' facial expressions in the process of online learning

4.2 Facial Expression Recognition and

Analysis

1000 pictures of students’ facial expressions were

input into our trained model for facial expression

recognition. The results are shown in Figure 5.

Figure 5. Analysis of facial expression recognition of 100

student volunteers

According to statistics, during online class,

students with a neutral expression ranked first. But it

is worth noting that among the seven expressions,

disgust emotion ranks second, and most of the

expressions appear in the middle and late stage of the

online studying. We speculate that this situation may

be related to the aversion of students due to the long

teaching time. In addition, anger, sadness, fear and

surprise account for a relatively small proportion,

indicating that students do not have large emotional

fluctuations during online classes.

After the class, we interviewed some volunteers.

In the interview, we learned that majority of students

have difficulty concentrating during distance

teaching, which will affect what they learn. In

addition, most students believe that virtual courses

usually take more time. When asked about the course

assignments completed this year compared with those

before COVID-19, nearly half said they spent more

time on these assignments.

4.3 SAS Prediction

We use facial expression recognition technology to

identify whether there are abnormalities in the

psychological state of volunteers. In order to verify

the accuracy of our experiment, we used the Self-

rating Anxiety Scale. SAS was compiled by William

W.K. Zung. The scale has become one of the most

commonly used psychological measurement tools for

psychologists.

The scale contains 20 subjective feeling items of

reaction anxiety. Each item is divided into four grades

according to the frequency of symptoms, from no or

little time to most or all of the time. Among them, 15

were positive scores and 5 were negative scores. As

shown in Table 2.

Prediction of Students’ Psychological States Based on Facial Expression Recognition Under the Inﬂuence of COVID-19

239

Table 2. Model accuracy

SN PROJECT

No or little

time

A small part

of the time

Quite a lot

of time

Most or all of

the time

I feel more nervous and anxious

than usual

1 2 3 4

2 I was scared for no reason 1 2 3 4

I tend to feel upset or frightened 1 2 3 4

I think I might be going crazy 1 2 3 4

Everything is fine and nothing

ad will ha

4 3 2 1

6 My hands and feet trembled 1 2 3 4

I suffer from headache, neck pain

and

ack pain

1 2 3 4

I feel weak and tired easily 1 2 3 4

I feel calm and easy to sit quietly 4 3 2 1

I feel my heart beating fast 1 2 3 4

I was distressed by bouts of

dizziness

1 2 3 4

I've had fainting attacks or feel

like I'll faint

1 2 3 4

13 It's easy for me to breathe 4 3 2 1

My hands and feet are numb and

tin

lin

1 2 3 4

I was afflicted with stomachache

and indi

estion

1 2 3 4

16 I often have to pee 1 2 3 4

My hands are often dry and warm 4 3 2 1

My face is red and hot 1 2 3 4

I sleep easily and sleep well all

night

4 3 2 1

20 I often have nightmares 1 2 3 4

Multiplying the total score by 1.23 to take an

integer is the standard score. The standard score

divides the severity of anxiety symptoms into none

(<50), mild (50-60), moderate (61-70) and severe (>

70).

According to the above standards, we made

statistics on the questionnaire test results of 100

volunteers, and the results are shown in the Figure 6.

According to statistics, 47% of the students have a

healthy mental state with no anxiety; 53% of the

students had different degrees of anxiety. Among

them, 38% of the students have mild anxiety, and the

standard indicators of some students are around the

critical line, which requires a long time of

psychological counseling and pressure release; 13%

of the students had moderate anxiety; 2% of the

students had severe anxiety.

Figure 6. Statistical chart of SAS scale test results

The psychological anxiety tendency of students

measured by SAS is about the same as that obtained

by facial expression recognition technology, which

shows that it is effective to use facial expression

recognition technology to predict students'

psychological problems.

NMDME 2022 - The International Conference on New Media Development and Modernized Education

240

4.4 Experiment Analysis

Anxiety is an autonomous response composed of

vague and uneasy feelings, which leads individuals to

experience or expect threatening or dangerous events

(Aloufi et al., 2021, Sun et al., 2016). Research shows

that appropriate anxiety can promote students'

learning. As shown in Figure 7. There is an inverted

U-shaped relationship between anxiety and students'

academic performance. Certain anxiety can make

students perform better, but once beyond a certain

stage, excessive anxiety will make students collapse

completely.

Figure 7. Anxiety-performance curve

For students, psychological symptoms of anxiety

include feeling nervous before class, panic, blank

brain during examination, feeling helpless when

doing homework, or showing interest in subjects

considered difficult, while physiological symptoms

include sweaty palms, accelerated heartbeat, or

dizziness and chest tightness (Ruffin et al., 2007).

Those with a high degree of anxiety will show

excessive tension and worry in terms of emotion,

including worries about the future. Through the

interview, we learned that students with severe

anxiety will worry about major accidents or bad

things in the future. Sometimes they will cry suddenly

at night and be extremely frightened. There will also

be physical manifestations, such as shortness of

breath, panic and other physical symptoms,

accompanied by serious sleep disorders, and the

possibility of suicide in serious cases. There are many

suicide cases of patients with severe anxiety disorder

in China every year.

Anxiety is one of the emotional behaviors that

affect students' daily life, such as learning, daily

activities and social networking. In the classroom

environment, anxiety disorder may be manifested in

students' behavior, like being easily irritable and

unable to focus on what teachers say. Deeper anxiety

is related to students' life and health. In addition, high

anxiety is also closely related to the poor performance

of low ability students (Whitaker Sena et al., 2007),

which will reduce memory and affect attention.

5 THE CAUSES OF STUDENTS'

LEARNING ANXIETY

Learning anxiety is a real phenomenon, so it is very

important to understand learning anxiety, especially

to be able to identify its source and deal with it.

Family status, living environment, learning

environment and psychological quality of college

students during the outbreak of COVID-19 affected

the mental health of college students by interviewing

the volunteers with different degrees of anxiety.

Through interviews, the reasons are as follows:

5.1 It is Difficult to Adapt to Abnormal

Learning State

In the most serious period of the epidemic, schools

were closed and students studied at home. Without a

collective learning atmosphere, students have

difficulty concentrating. Teachers use a large number

of assignments to urge students to learn during online

studying. Some students need to spend twice as much

time completing their homework as most students in

school, which makes students feel bored (Zhu et al.,

2020). The efficiency of students studying at home is

far less than that in school. Most students are annoyed

by wasting time. Such prolonged repetitive cycle for

a long time has led to anxiety or depression (Cao et

al., 2020).

5.2 Poor Environmental Adaptability

With the resumption of college and students returning

to school, universities implement semi closed

management to reduce students' unnecessary going

out. Students need to report in advance before

entering or leaving the school and measure

temperature at the gate of the campus. Under this

circumstance, students' daily shopping, traveling,

social practice and other social activities have

decreased sharply, making students feel strange,

repressive and monitored (Kılınçel et al., 2021).

These conditions lead to the decline of students'

psychological adaptability and the disorder of

psychological adjustment ability, which affect their

normal study and life in school (Zawadka et al.,

2021).

Prediction of Students’ Psychological States Based on Facial Expression Recognition Under the Inﬂuence of COVID-19

241

5.3 Family Factors

In the early stage of the epidemic, due to the need for

epidemic prevention and isolation, teachers and

students taught and studied at home through online

classes. Students study and live at home with their

families for a long time. They are prone to quarrels

and contradictions due to family chores, tense parent-

child relations and disharmonious family relations,

which can lead to certain harm and impact on

students' mental health (Li et al., 2021). At the same

time, due to the local epidemic situation, some

students' families are isolated, which makes their

parents cannot go out to work, and their family

economic income decreases sharply. The burden of

family life has brought invisible psychological

pressure and psychological anxiety to the students

themselves.

According to the research of this paper, students

need some intervention to reduce anxiety and help

improve academic performance (Chen G et al., 2020).

Learning anxiety intervention aims to help students

deal with anxiety problems in the process of learning

(Deng et al., 2021).

6 CONCLUSION

This study used facial expression recognition

technology to predict students' psychological

problems. The latest information provided of this

research results offers a necessary basis for the

research and intervention of College Students'

psychological problems during COVID-19. It is

worth noting that COVID-19 has indeed brought

different psychological effects to students. This paper

emphasizes that the relationship between the impact

of the epidemic and different degrees of anxiety

should be analyzed longitudinally, and intervention

measures should be taken from a scientific

perspective to eliminate the fear of students in

educational activities.

ACKNOWLEDGEMENTS

This work was supported National Natural Science

Foundation of China (61373148), National Social

Science Fund of China (12BXW040); Science

Foundation of Ministry of Education of

China(14YJC860042), Shandong Provincial Social

Science Planning Project

(18CxWJ01,18BJYJ04,19BJCJ51).

REFERENCES

Alzahrani, L., Seth, K.P. (2021). Factors influencing

students’ satisfaction with continuous use of learning

management systems during the covid-19 pandemic:

An empirical study. Education and Information

Technologies, 1–19.

Aloufi, M.A., Jarden, R.J., Gerdtz, M., Kapp, S. (2021).

Reducing stress, anxiety and depression in

undergraduate nursing students: Systematic review.

Nurse Education Today, 104877.

Bartlett, M.S., Littlewort, G., Fasel, I., Movellan, J.R.

(2003). Real time face detection and facial expression

recognition: development and applications to human

computer interaction. In: 2003 Conference on

Computer Vision and Pattern Recognition Workshop,

vol. 5(pp. 53–53). IEEE.

Căleanu, C.-D. (2013). Face expression recognition: A brief

overview of the last decade. In: 2013 IEEE 8th

International Symposium on Applied Computational

Intelligence and Informatics (SACI) (pp. 157–161).

IEEE.

Chanchal, A.K., Dutta, M., Scholar, M.E., NITTTR,

Electronics, H., Deptt, C. (2016). Recognition of

emotions from facial expressions and its application in

car driving system. International Journal of Advanced

Research in Electronics and Communication

Engineering 5(10), 2398–2406.

Cao, W., Fang, Z., Hou, G., Han, M., Xu, X., Dong, J.,

Zheng, J. (2020). The psychological impact of the

covid-19 epidemic on college students in China.

Psychiatry research 287, 112934.

Chen G, Zhang J, Chan C K K, et al. The link between

student‐perceived teacher talk and student enjoyment,

anxiety and discursive engagement in the classroom[J].

British Educational Research Journal, 2020, 46(3): 631-

652.

Calbi, M., Langiulli, N., Ferroni, F., Montalti, M.,

Kolesnikov, A., Gallese, V., Umiltà, M.A. (2021). The

consequences of covid-19 on social interactions: an

online study on face covering. Scientific reports 11(1),

1–10.

Deng J, Zhou F, Hou W, et al. The prevalence of depressive

symptoms, anxiety symptoms and sleep disturbance in

higher education students during the COVID-19

pandemic: A systematic review and meta-analysis[J].

Psychiatry Research, 2021, 301: 113863.

Fu, W., Yan, S., Zong, Q., Anderson-Luxford, D., Song, X.,

Lv, Z., Lv, C. (2021). Mental health of college students

during the covid-19 epidemic in china. Journal of

Affective Disorders 280, 7–10.

Ilyas, C.M.A., Nunes, R., Nasrollahi, K., Rehm, M.,

Moeslund, T.B. (2021). Deep emotion recognition

through upper body movements and facial expression.

In: VISIGRAPP (5: VISAPP), pp. 669–679.

Jabon, M., Bailenson, J., Pontikakis, E., Takayama, L.,

Nass, C. (2010). Facial expression analysis for

predicting unsafe driving behavior. IEEE Pervasive

Computing 10(4), 84–95.

NMDME 2022 - The International Conference on New Media Development and Modernized Education

242

Kumar, G.R., Kumar, R.K., Sanyal, G. (2017). Facial

emotion analysis using deep convolution neural

network. In: 2017 International Conference on Signal

Processing and Communication (ICSPC) (pp. 369–

374). IEEE.

Kılınçel, Ş., Kılınçel, O., Muratdağı, G., Aydın, A., Usta,

M.B. (2021). Factors affecting the anxiety levels of

adolescents in home-quarantine during covid-19

pandemic in turkey. Asia-Pacific Psychiatry 13(2),

12406.

Luqin, S. (2019). A survey of facial expression recognition

based on convolutional neural network. In: 2019

IEEE/ACIS 18th International Conference on

Computer and Information Science (ICIS) (pp. 1–6).

IEEE.

Li, W., Zhang, Y., Wang, J., Ozaki, A., Wang, Q., Chen,

Y., Jiang, Q. (2021). Association of home quarantine

and mental health among teenagers in Wuhan, China,

during the covid-19 pandemic. JAMA pediatrics

175(3), 313–316.

Li, Z., Zhang, T., Jing, X., Wang, Y. (2021). Facial

expression-based analysis on emotion correlations,

hotspots, and potential occurrence of urban crimes.

Alexandria Engineering Journal 60(1), 1411–1420.

Mohiadeen, A., et al. (2021). Evaluating students’ reaction

to lectures using facial expression recognition. Turkish

Journal of Computer and Mathematics Education

(TURCOMAT) 12(3), 1952–1956.

Panskyi, T., Korzeniewska, E., Serwach, M., Grudzień, K.

(2021). New realities for polish primary school

informatics education affected by covid-19. Education

and Information Technologies, 1–28.

Ruffin, P. (2007). A real fear: It’s more than stage fright,

math anxiety can derail academic or professional

success, but some scholars are working to help students

get over it. Diverse Issues in Higher Education.

http://findarticles.

com/p/articles/mi_m0WMX/is_2_24/ai_n18744928,

2007.

Sun, F.-K., Long, A., Tseng, Y.S., Huang, H.-M., You, J.-

H., Chiang, C.-Y. (2016). Undergraduate student

nurses’ lived experiences of anxiety during their first

clinical practicum: A phenomenological study. Nurse

education today 37, 21–26.

Tulaskar, R., Turunen, M. (2021). What students want?

experiences, challenges, and engagement during

emergency remote learning amidst covid-19 crisis.

Education and information technologies, 1–37.

Whitaker Sena, J.D., Lowe, P.A., Lee, S.W. (2007).

Significant predictors of test anxiety among students

with and without learning disabilities. Journal of

learning disabilities 40(4), 360–376.

Zhao, L., Wang, Z., Wang, X., Liu, Q. (2018). Driver

drowsiness detection using facial dynamic fusion

information and a dbn. IET Intelligent Transport

Systems 12(2), 127–133.

Zhu, T. (2020). Empirical research on the application of

online teaching in Chinese colleges and universities

under the situation of novel coronavirus pneumonia

prevention and control. International Journal of

Emerging Technologies in Learning (iJET) 15(11),

119–136.

Zawadka, J., Miękisz, A., Nowakowska, I., Plewko, J.,

Kochańska, M., Haman, E. (2021). Remote learning

among students with and without reading difficulties

during the initial stages of the covid-19 pandemic.

Education and Information Technologies, 1–22.

Prediction of Students’ Psychological States Based on Facial Expression Recognition Under the Inﬂuence of COVID-19

243