University 2.0 Introduction: Emotion is the expression of

 

 

 

University of
Bahrain

College of
Information Technology

Best services for writing your paper according to Trustpilot

Premium Partner
From $18.00 per page
4,8 / 5
4,80
Writers Experience
4,80
Delivery
4,90
Support
4,70
Price
Recommended Service
From $13.90 per page
4,6 / 5
4,70
Writers Experience
4,70
Delivery
4,60
Support
4,60
Price
From $20.00 per page
4,5 / 5
4,80
Writers Experience
4,50
Delivery
4,40
Support
4,10
Price
* All Partners were chosen among 50+ writing services by our Customer Satisfaction Team

IT 603

 

 

 

 

Assignment5

Paper Critique

 

Article:

 (Emotion recognition based on EEG features in
movie clips with channel selection)

 

 

 

 

 

 

Prepared by:

Student name: Rabab Hameed Ali Alrahim

ID.:20170027

Program: MSc. In Information Technology

Table of content

 

1.0Objective …………….………………………………………….…… 3

2.0Introduction …………………………………………………………. 3

3.0Paper background …………………………………….……………. 3

4.0Paper summary ………………………………………………………. 3

5.0Critique of Paper’s abstract………………………….…..…………. 4

6.0Critique of Paper’s Keywords………………………………….…… 4

7.0Critique of Paper’s Introduction………………………………………4

          7.1Background………………………………………….…….……
4

          7.2Literature
review………………………………………..………. 5

8.0Critique of Paper’s Materials…………………………………….… 5

          8.1Database and
participants questionnaire………………………..5

          8. 2Stimulus
material……………..…………………..……………6

          8.3Task……………..……………………………………………………6

          8.4Participants
ratings…………………………..…………………6

          8.5EEG signal
recordings…………………..………………………6

9.0Critique of Paper’s Method……………………………….…………6

10.0Critique of Paper’s Experimental results………….………………7

11.0Critique of Paper’s Results and discussion…….…………………7

12.0Critique of Paper’s Conclusion……………………………………8

13.0 Other Published Work in the Same Area…………………………8

14.0Conclusion……………..……………………………………….……8

15.0References……………..………………………………….…………9

 

 

 

1.0 Objective:

·       
This assignment is to critique (Emotion recognition
based on EEG features in movie clips with channel selection) article and
discuss how good or bad is it.

 

2.0 Introduction:

Emotion is the expression of human feelings such
as love, hate, anger, fear etc. It is the reaction against a specific event.
Emotion can tell a lot about person and it can explain what words can’t do, it
shows what person may hide and because of that psychologist and scientist focus
in emotion recognition by using different methods to understand human and their
feelings more. One of these methods is using EEG signal for emotion recognition,
“Emotion
recognition based on EEG features in movie clips with channel selection” 1
is an article from Brain Informatics journal which published by Springer
talking about this method.

As it defined in The Oxford Dictionary “Critique is to evaluate (a theory or practice) in a detailed
and analytical way.” 2 In the following pages I am going to critique this
article and analysis it in details and discuss everything from the structure of
the article to the depth of the subject and the way it presented.

I chose this article because it is one of my
interest topics as a biomedical engineer. Also, it is well structure and can
fulfils the assignment requirement.

 

3.0 Paper background:

This paper prepared by two authors, the
first one is Mehmet Siraç Özerdem who is one of an Electrical and Electronics
Engineering department of Dicle University in Diyarbak?r, Turkey and the other
one is Hasan Polat who is also one of Electrical and Electronics Engineering
department but of Mus Alparslan University in Mu?, Turkey.

Özerdem and Polat work in different work
based on EEG signals, and they wrote several paper together in this topics,
such as “The comparison of wavelet and empirical mode decomposition method
in prediction of sleep stages from EEG signals” 3 published on September
2017, and “Familiarity effect of emotional
stimuli onto EEG signals” 4 published on October 2016.

 

4.0 Paper summary:

The paper is discussing a study on emotion
recognition based on EEG signals. The experiment applied on 32 participants and
held in two universities (Twente and Geneva). Music clips used to stimulate
emotions and EEG signals record during listening to music. EEG channels for
participants evaluated separately and 5 channels having the highest performance
were selected. The final feature vectors of the emotions classify using two
methods multilayer perceptron neural network (MLPNN) and k-nearest neighborhood
(kNN) and the classification performance achieved, computed and compare at the
end of the study experiment.

 

5.0 Critique of Paper’s
abstract:

The abstract of the article is well-structured
and clear. Its size good and there is no difficult words or uneasy to
understand. Authors started the abstract by mentioning the importance of emotion,
they state that brain–computer interface (BCI) system is not compatible with
emotion recognition and declare how EEG signals involve in emotion recognition
and what are the advantages of it to link it to the paper topic to talk about
the methods used in this study. The last part of the abstract was explaining
how the final result is achieved.

 

6.0 Critique of Paper’s
Keywords:

The chosen keywords are good but it’s
better if ‘Emotion recognition’ was mentioned.

 

7.0 Critique of Paper’s
Introduction:

7.1 Background

The authors begin by introducing the area
which the work is related to, Emotion and EEG signals. Start by declaring why
emotion is important then identifying EEG signal and linked this to the study
by specifying that it is the most used technique for emotion recognition.

In the third paragraph previous method use
for emotion recognition (EEG-based BCI system) clarified with the advantages
and disadvantages of it and the future effect of interpretation people emotion
using this system. Then follow up by the two ways use for discriminating
emotion.

The flow of the ideas and its sequence is
very good and clear and it’s presented in an organized structure which allows
people from different specialization to read it. However, authors didn’t take
concern some point for example in the third paragraph SAM acronym is not
defined.

 

 

7.2 Literature review

The literature review combined with the
introduction and authors start it by stating the three areas the studies
utilizes. The first two area is not related to the EEG-based emotion
recognition hence the authors didn’t explain them, the focus has been on the
literature which is used EEG-based method and this is proper decision, thus
time will not be wasted on something that’s not related to work. Also, the
person which going to read the paper will not be distracted with a lot of
information not in the main topic but if he interested on it he can search by
himself.

In the second paragraph of literature the
studies used EEG signals in health, games and advertisement areas were
mentioned.

In the followed paragraph 5 main studies
related to emotion recognition using EEG signal approach was listed. Authors
briefly clarify the number of channels used, methods used and the percentages
of accuracy of the result achieved. Then mention the limitation on the previous
studies which were (limited number of channels used).

The references of the literature were
written accurately in the reference section and the latest publication studies
were used.

By the end of the introduction section
authors state the aim of the study they are going to conduct which will differ
from other studies by using preprocessing of channel selection.  Then briefly talk about the process which is
going to be followed, method will use and how the result going to be fulfilled.

Last paragraph of the introduction authors
state the structure of the paper and in which form it organizes, this a good
point as it is going to give the reader idea about the paper as it works same
as a table of content.

 

8.0Critique of Paper’s
Materials:

8.1Database and participants questionnaire

The study uses DEAP database and the
authors didn’t define this acronym and 32 participant (15 female and 17 male)
in the age (between 23and 37) took part in the experiment. The experiment held
in two universities (Twente and Geneva). There was some difference between the
two universities, because of that the group participated and examined in Twente
university used in the experiment. However, authors didn’t state why they used
this group not the other one and also didn’t state why they use this range of
age but in my point of view I think they use this age range because it easier
to recognize the emotion as the participant is young.

8. 2Stimulus material

For stimulating emotion in the participant
to record the DEAP database authors used music. And this is a good choose
because music produces different feelings and that can give different emotion.

120 music clips were select initially semi automatically
and manually and after testing only 40 from them determined to be to extract
emotion for the final experiment. However, authors didn’t provide reasons and
the parameters that they follow to choose the 40 music clips.

8.3Task

The authors clarify the task and steps used
in a very precise and clear way. They explain the paradigm which followed to
record EEG for each participant and mention the steps used in the paradigm and
illustrate it using a diagram.

8.4Participants ratings

During the experiment participant was given
assessment to do after each music clip to state their emotional.

The authors specify the assessment used
which is (SAM) and explain how participant rate and mention that their
evaluation is taken into consideration in the study.

Authors provide figure to show the SAM
scale and another figure show the EEG signal related to negative and positive
emotion and that support their explanation and make it more understandable and
clear to the reader.

8.5EEG signal recordings

In this section authors briefly and
properly explain how the EEG signals recorded and using what, as well as how to
extract it and remove unwanted part of the signal such as artifact.

 

9.0Critique of Paper’s Method:

In the method section authors first give
general idea about the methods used in the study then discuss in details about
each in 4 separate sections. The ideas has smoothly flow with well written,
organized structure.

First authors describe DWS method which is
used for feature extraction from EEG signals. Method explained generally and followed
by mathematical definition for signal decomposition, supported by figure shows
signal decomposition. Frequency ranges related to DWT shows in a table and
authors refer to a literature to decide which frequency to choose for
recognizing different emotion.

Secondly, explained which band of wavelet
coefficients after decomposition of signals were used and listed the parameters
selected to decrease the dimensionality of feature vectors.

Thirdly, they explain the two
classification methods used for classifying negative and positive emotion in
details. Each algorithm explain at the beginning then describe how it applied
in the study with mathematical definitions presented in the paper. This way of
presentation this part of study is an excellent way to show the reader the full
picture of what is done, how and why.

Finally, Authors state the 3 performance
criteria used in the study and how they get them by mathematical equations.

 

10.0Critique of Paper’s Experimental
results:

Experimental results explain in 3 sections.
In the first section authors state that after examine 32 channels for each
participant 5 of them were chosen and explain how they determined and present
the final result by a graph shown the EEG channels which have the best
classification performances. In this section some sentence were repeated.

In the second and third sections authors
explain how final feature vectors obtained and the result of classification
which applied for each participant using  MLPNN and kNN algorithm were provided in a tables with the averages, and the
comparison of them shown in a bar chart. The results from both algorithms
quietly similar.

 

11.0Critique of Paper’s Results
and discussion:

The results and discussion showed that the
study objective was met. The results were presented in list format with brief
discussion and that make it for easy the reader to go through it.

The result in simple form is as authors
declare as follow:

·       
Wavelet coefficients obtained using DWT and evaluated
as feature vectors.

·       
5 channels having the best classification performances
and they used for the study

·       
EEG signals classified using MLPNN and kNN methods

·       
Results obtained using MLPNN and kNN methods and compare between them.

·       
States that MLPNN and kNN which used in the study give
results for classification of emotions.

 

12.0Critique of Paper’s Conclusion:

The conclusion is briefed, well written
summarized the main point of the study.

It was started stating the aim of the study
and summarize what done in the study and result got. And end by indicating what
can be done in the future and what can be done to increase the classification
success.

 

13.0 Other Published Work in
the Same Area:

By searching in Google scholar and other
search engine I get some papers in the same area and I chose the most similar
to compare it with the paper that I critiqued.

Murugappan Murugappan, Nagarajan
Ramachandran, Yaacob Sazali “Classification of human emotion from EEG
using discrete wavelet transform”5, J. Biomedical Science and
Engineering, 2010, 3, 390-396.

This paper discusses classification of EEG
signal for emotion recognition using wavelet transforms same as the paper which
has been critique. However, authors in this paper introduce few different
methods. The study was
examined on 20 participant (3 female and 17 male) age between 21-39 years. They
use Surface Laplacian (SL) filtering method for preprocessing EEG signals and
used Linear Discriminant Analysis (LDA) with kNN methods for classification
emotion rather than MLPNN methods. For stimulating emotion they used Video
clips instead of music clips.

 

14.0Conclusion:

This assignment was done to critique a
paper in any topic related to IT. My choice was a paper written by Mehmet
Sirac¸ O¨ zerdem and Hasan Polat, with the following title “Emotion recognition based on EEG features in movie clips with
channel selection”.

The paper discusses study on emotion
recognition using EEG signals and it was applied on thirty-two healthy
participant. The emotion has been classified using channel selection by
applying to methods MLPNN and kNN.

My overall evaluation is that it is a good
paper well written and has a good structure. The flow of ideas and their
explanation was excellent and the references used are wide and enough to
support the study.  The presentation of
ideas and methods used was well explained.

 

15.0References:

1  Özerdem, M. S., & Polat, H. (2017).
Emotion recognition based on EEG features in movie clips with channel
selection. Brain Informatics, 4(4), 241-252. doi:10.1007/s40708-017-0069-3

2  Critique | Definition of critique in
English by Oxford Dictionaries. (n.d.). Retrieved December 02, 2017, from https://en.oxforddictionaries.com/definition/critique

3   Özerdem,
M. S., Polat, H. & Akin.M .The comparison of wavelet and empirical mode
decomposition method in prediction of sleep stages from EEG signals.  Artificial Intelligence and Data Processing
Symposium (IDAP), 2017 International. 10.1109/IDAP.2017.8090253.IEEE.

4  Özerdem, M. S.,& Polat, H. Familiarity
effect of emotional stimuli onto EEG signals. n: Medical Technologies National Congress (TIPTEKNO),
2016. 10.1109/TIPTEKNO.2016.7863119.IEEE.

5  Murugappan, M. , Ramachandran, N. and
Sazali, Y. (2010) Classification of human emotion from EEG using discrete
wavelet transform. Journal of Biomedical Science and Engineering, 3, 390-396.
doi: 10.4236/jbise.2010.34054.