Artificial Intelligence and Algorithms in

Intelligent Systems Radek Silhavy pdf download

https://textbookfull.com/product/artificial-intelligence-and-algorithms-in-intelligent-systems-
radek-silhavy/

★★★★★ 4.7/5.0 (23 reviews) ✓ 86 downloads ■ TOP RATED

"Perfect download, no issues at all. Highly recommend!" - Mike D.

DOWNLOAD EBOOK
 Artificial Intelligence and Algorithms in Intelligent
Systems Radek Silhavy pdf download

TEXTBOOK EBOOK TEXTBOOK FULL

Available Formats

■ PDF eBook Study Guide TextBook

EXCLUSIVE 2025 EDUCATIONAL COLLECTION - LIMITED TIME

INSTANT DOWNLOAD VIEW LIBRARY

Collection Highlights

Cybernetics and Algorithms in Intelligent Systems Radek

Silhavy

Software Engineering and Algorithms in Intelligent Systems

Radek Silhavy

Artificial Intelligence Methods in Intelligent Algorithms

Proceedings of 8th Computer Science On line Conference
2019 Vol 2 Radek Silhavy

Computational and Statistical Methods in Intelligent

Systems Radek Silhavy
Intelligent Systems in Cybernetics and Automation Control
Theory Radek Silhavy

Software Engineering Perspectives in Intelligent Systems:

Proceedings of 4th Computational Methods in Systems and
Software 2020, Vol.1 Radek Silhavy

Software Engineering Perspectives in Intelligent Systems:

Proceedings of 4th Computational Methods in Systems and
Software 2020, Vol.2 Radek Silhavy

Intelligent Algorithms in Software Engineering:

Proceedings of the 9th Computer Science On-line Conference
2020, Volume 1 Radek Silhavy

Computational Statistics and Mathematical Modeling Methods

in Intelligent Systems Proceedings of 3rd Computational
Methods in Systems and Software 2019 Vol 2 Radek Silhavy
Advances in Intelligent Systems and Computing 764

Radek Silhavy Editor

Artificial
Intelligence and
Algorithms in
Intelligent Systems
Proceedings of 7th Computer Science
On-line Conference 2018, Volume 2
Advances in Intelligent Systems and Computing

Volume 764

Series editor
Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland
e-mail: [email protected]
The series “Advances in Intelligent Systems and Computing” contains publications on theory,
applications, and design methods of Intelligent Systems and Intelligent Computing. Virtually all
disciplines such as engineering, natural sciences, computer and information science, ICT, economics,
business, e-commerce, environment, healthcare, life science are covered. The list of topics spans all the
areas of modern intelligent systems and computing such as: computational intelligence, soft computing
including neural networks, fuzzy systems, evolutionary computing and the fusion of these paradigms,
social intelligence, ambient intelligence, computational neuroscience, artificial life, virtual worlds and
society, cognitive science and systems, Perception and Vision, DNA and immune based systems,
self-organizing and adaptive systems, e-Learning and teaching, human-centered and human-centric
computing, recommender systems, intelligent control, robotics and mechatronics including
human-machine teaming, knowledge-based paradigms, learning paradigms, machine ethics, intelligent
data analysis, knowledge management, intelligent agents, intelligent decision making and support,
intelligent network security, trust management, interactive entertainment, Web intelligence and multimedia.
The publications within “Advances in Intelligent Systems and Computing” are primarily proceedings
of important conferences, symposia and congresses. They cover significant recent developments in the
field, both of a foundational and applicable character. An important characteristic feature of the series is
the short publication time and world-wide distribution. This permits a rapid and broad dissemination of
research results.

Advisory Board
Chairman
Nikhil R. Pal, Indian Statistical Institute, Kolkata, India
e-mail: [email protected]
Members
Rafael Bello Perez, Universidad Central “Marta Abreu” de Las Villas, Santa Clara, Cuba
e-mail: [email protected]
Emilio S. Corchado, University of Salamanca, Salamanca, Spain
e-mail: [email protected]
Hani Hagras, University of Essex, Colchester, UK
e-mail: [email protected]
László T. Kóczy, Széchenyi István University, Győr, Hungary
e-mail: [email protected]
Vladik Kreinovich, University of Texas at El Paso, El Paso, USA
e-mail: [email protected]
Chin-Teng Lin, National Chiao Tung University, Hsinchu, Taiwan
e-mail: [email protected]
Jie Lu, University of Technology, Sydney, Australia
e-mail: [email protected]
Patricia Melin, Tijuana Institute of Technology, Tijuana, Mexico
e-mail: [email protected]
Nadia Nedjah, State University of Rio de Janeiro, Rio de Janeiro, Brazil
e-mail: [email protected]
Ngoc Thanh Nguyen, Wroclaw University of Technology, Wroclaw, Poland
e-mail: [email protected]
Jun Wang, The Chinese University of Hong Kong, Shatin, Hong Kong
e-mail: [email protected]

More information about this series at http://www.springer.com/series/11156

Radek Silhavy
Editor

Artificial Intelligence
and Algorithms in Intelligent
Systems
Proceedings of 7th Computer Science On-line
Conference 2018, Volume 2

123
Editor
Radek Silhavy
Faculty of Applied Informatics
Tomas Bata University in Zlín
Zlín
Czech Republic

ISSN 2194-5357 ISSN 2194-5365 (electronic)

Advances in Intelligent Systems and Computing
ISBN 978-3-319-91188-5 ISBN 978-3-319-91189-2 (eBook)
https://doi.org/10.1007/978-3-319-91189-2
Library of Congress Control Number: 2018943432

© Springer International Publishing AG, part of Springer Nature 2019

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.

Printed on acid-free paper

This Springer imprint is published by the registered company Springer International Publishing AG
part of Springer Nature
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface

This book constitutes the refereed proceedings of the modern trends and approaches
of artificial intelligence research and its application to intelligent systems. The paper
discusses hybridization of algorithms, new trends in neural networks, optimization
algorithms, and real-life issues related to artificial methods’ application.
This book constitutes the refereed proceedings of the Artificial Intelligence and
Algorithms in Intelligent Systems of the 7th Computer Science On-line Conference
2018 (CSOC 2018), held online in April 2018.
CSOC 2018 has received (all sections) 265 submissions, 141 of them were
accepted for publication. More than 60% of accepted submissions were received
from Europe, 30% from Asia, 5% from Africa, and 5% from America. Researchers
from 30 countries participated in CSOC 2018 conference.
CSOC 2018 conference intends to provide an international forum for the dis-
cussion of the latest high-quality research results in all areas related to Computer
Science. The addressed topics are the theoretical aspects and applications of
Computer Science, Artificial Intelligences, Cybernetics, Automation Control
Theory and Software Engineering.
Computer Science On-line Conference is held online and modern communica-
tion technology, which is broadly used to improve the traditional concept of sci-
entific conferences. It brings equal opportunity to participate to all researchers
around the world.
I believe that you will find following proceedings interesting and useful for your
own research work.

March 2018 Radek Silhavy

Editor

v
Organization

Program Committee
Program Committee Chairs

Petr Silhavy Tomas Bata University in Zlin, Faculty of Applied

Informatics
Radek Silhavy Tomas Bata University in Zlin, Faculty of Applied
Informatics
Zdenka Prokopova Tomas Bata University in Zlin, Faculty of Applied
Informatics
Roman Senkerik Tomas Bata University in Zlin, Faculty of Applied
Informatics
Roman Prokop Tomas Bata University in Zlin, Faculty of Applied
Informatics
Viacheslav Zelentsov Doctor of Engineering Sciences, Chief Researcher
of St. Petersburg Institute for Informatics
and Automation of Russian Academy of Sciences
(SPIIRAS)

Program Committee Members

Boguslaw Cyganek Department of Computer Science, University

of Science and Technology, Krakow, Poland
Krzysztof Okarma Faculty of Electrical Engineering, West Pomeranian
University of Technology, Szczecin, Poland
Monika Bakosova Institute of Information Engineering, Automation
and Mathematics, Slovak University of Technology,
Bratislava, Slovak Republic

vii
viii Organization

Pavel Vaclavek Faculty of Electrical Engineering and Communication,

Brno University of Technology, Brno,
Czech Republic
Miroslaw Ochodek Faculty of Computing, Poznan University
of Technology, Poznan, Poland
Olga Brovkina Global Change Research Centre Academy of Science
of the Czech Republic, Brno, Czech Republic
and Mendel University of Brno, Czech Republic
Elarbi Badidi College of Information Technology, United Arab
Emirates University, Al Ain, United Arab Emirates
Luis Alberto Morales Head of the Master Program in Computer Science,
Rosales Superior Technological Institute of Misantla,
Mexico
Mariana Lobato Baes Superior Technological of Libres, Mexico
Abdessattar Chaâri Laboratory of Sciences and Techniques of Automatic
control and Computer engineering, University
of Sfax, Tunisian Republic
Gopal Sakarkar Shri. Ramdeobaba College of Engineering
and Management, Republic of India
V. V. Krishna GD Rungta College of Engineering & Technology,
Maddinala Republic of India
Anand N. Khobragade Maharashtra Remote Sensing Applications Centre,
Republic of India
Abdallah Handoura Computer and Communication Laboratory, Telecom
Bretagne, France

Technical Program Committee Members

Ivo Bukovsky Roman Senkerik

Maciej Majewski Petr Silhavy
Miroslaw Ochodek Radek Silhavy
Bronislav Chramcov Jiri Vojtesek
Eric Afful Dazie Eva Volna
Michal Bliznak Janez Brest
Donald Davendra Ales Zamuda
Radim Farana Roman Prokop
Martin Kotyrba Boguslaw Cyganek
Erik Kral Krzysztof Okarma
David Malanik Monika Bakosova
Michal Pluhacek Pavel Vaclavek
Zdenka Prokopova Olga Brovkina
Martin Sysel Elarbi Badidi
Organization ix

Organizing Committee Chair

Radek Silhavy Tomas Bata University in Zlin, Faculty of Applied

Informatics

Conference Organizer (Production)

OpenPublish.eu s.r.o.
Web: http://www.openpublish.eu
Email: [email protected]

Conference Website, Call for Papers

http://www.openpublish.eu
Contents

Analysis of Affective and Gender Factors in Image Comprehension

of Visual Advertisement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Gabrielė Liaudanskaitė, Gabrielė Saulytė, Julijus Jakutavičius,
Eglė Vaičiukynaitė, Ligita Zailskaitė-Jakštė, and Robertas Damaševičius
FARIP: Framework for Artifact Removal for Image Processing
Using JPEG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
T. M. Shashidhar and K. B. Ramesh
SOPA: Search Optimization Based Predictive Approach for Design
Optimization in FinFET/SRAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
H. Girish and D. R. Shashikumar
Analysis of the Quality of the Painting Process Using Preprocessing
Techniques of Text Mining . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Veronika Simoncicova, Pavol Tanuska, Hans-Christian Heidecke,
and Stefan Rydzi
Bioinspired Algorithm for 2D Packing Problem . . . . . . . . . . . . . . . . . . . 39
Vladimir Kureichik, Liliya Kureichik, Vladimir Kureichik, Jr.,
and Daria Zaruba
Authorship Identification Using Random Projections . . . . . . . . . . . . . . . 47
Robertas Damaševičius, Jurgita Kapočiūtė-Dzikienė, and Marcin Woźniak
A Method for Intelligent Quality Assessment of a Gearbox Using
Antipatterns and Convolutional Neural Networks . . . . . . . . . . . . . . . . . 57
Andrzej Tuchołka, Maciej Majewski, Wojciech Kacalak,
and Zbigniew Budniak
Spark-Based Classification Algorithms for Daily Living Activities . . . . . 69
Dorin Moldovan, Marcel Antal, Claudia Pop, Adrian Olosutean,
Tudor Cioara, Ionut Anghel, and Ioan Salomie

xi
xii Contents

Fast Adaptive Image Binarization Using the Region

Based Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
Hubert Michalak and Krzysztof Okarma
Semantic Query Suggestion Based on Optimized
Random Forests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
Aytuğ Onan
Financial Knowledge Instantiation from Semi-structured,
Heterogeneous Data Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
Francisco García-Sánchez, José Antonio García-Díaz,
Juan Miguel Gómez-Berbís, and Rafael Valencia-García
Hierarchical Fuzzy Deep Leaning Networks for Predicting Human
Behavior in Strategic Setups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
Arindam Chaudhuri and Soumya K. Ghosh
Fuzzy-Expert System for Customer Behavior Prediction . . . . . . . . . . . . 122
Monika Frankeová, Radim Farana, Ivo Formánek, and Bogdan Walek
A Binary Grasshopper Algorithm Applied
to the Knapsack Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
Hernan Pinto, Alvaro Peña, Matías Valenzuela, and Andrés Fernández
Artificial Neural Networks Implementing Maximum Likelihood
Estimator for Passive Radars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
Timofey Shevgunov and Evgeniy Efimov
Using Query Expansion for Cross-Lingual Mathematical
Terminology Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 154
Velislava Stoykova and Ranka Stankovic
Text Summarization Techniques for Meta Description Generation
in Process of Search Engine Optimization . . . . . . . . . . . . . . . . . . . . . . . 165
Goran Matošević
Integration of Models of Adaptive Behavior of Ant and Bee Colony . . . 174
Boris K. Lebedev, Oleg B. Lebedev, Elena M. Lebedeva,
and Andrey I. Kostyuk
Optimization of Multistage Tourist Route for Electric Vehicle . . . . . . . . 186
Joanna Karbowska-Chilinska and Kacper Chociej
Enhancing Stratified Graph Sampling Algorithms Based
on Approximate Degree Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
Junpeng Zhu, Hui Li, Mei Chen, Zhenyu Dai, and Ming Zhu
MIC-KMeans: A Maximum Information Coefficient Based
High-Dimensional Clustering Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . 208
Ruping Wang, Hui Li, Mei Chen, Zhenyu Dai, and Ming Zhu
Contents xiii

DACC: A Data Exploration Method for High-Dimensional

Data Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
Qingnan Zhao, Hui Li, Mei Chen, Zhenyu Dai, and Ming Zhu
Multi-targets Tracking of Multiple Instance Boosting Combining
with Particle Filtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 230
Hongxia Chu, Kejun Wang, Yumin Han, Rongyi Zhang, and Xifeng Wang
An Enhance Approach of Filtering to Select Adaptive IMFs of EEMD
in Fiber Optic Sensor for Oxidized Carbon Steel . . . . . . . . . . . . . . . . . . 241
Nur Syakirah Mohd Jaafar, Izzatdin Abdul Aziz, Jafreezal Jaafar,
Ahmad Kamil Mahmood, and Abdul Rehman Gilal
Hyper-heuristical Particle Swarm Method
for MR Images Segmentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
Samer El-Khatib, Yuri Skobtsov, Sergey Rodzin,
and Viacheslav Zelentsov
A Hybrid SAE and CNN Classifier for Motor Imagery
EEG Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
Xianlun Tang, Jiwei Yang, and Hui Wan
Semantic Bookmark System for Dynamic Modeling
of Users Browsing Preferences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
Syed Khurram Ali Shah, Shah Khusro, Irfan Ullah,
and Muhammad Abid Khan
Models, Algorithms and Monitoring System of the Technical
Condition of the Launch Vehicle “Soyuz-2” at All Stages
of Its Life Cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 288
Aleksey D. Bakhmut, Кljucharjov A. Alexander, Aleksey V. Krylov,
Michael Yu. Okhtilev, Pavel A. Okhtilev, Anton V. Ustinov,
and Alexander E. Zyanchurin
Proactive Management of Complex Objects Using
Precedent Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 298
Aleksey D. Bakhmut, Aleksey V. Krylov, Margaret A. Krylova,
Michael Yu. Okhtilev, Pavel A. Okhtilev, and Boris V. Sokolov
Artificial Intelligence and Algorithms in Intelligent Systems . . . . . . . . . 308
Carla Sofia R. Silva and Jose Manuel Fonseca
Hierarchical System for Evaluating Professional Competencies Using
Takagi-Sugeno Rules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 318
Ondrej Pektor, Bogdan Walek, Ivo Martinik, and Michal Jaluvka
xiv Contents

Discovering Association Rules of Information Dissemination About

Geoinformatics University Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 326
Zdena Dobesova
Predicting User Age by Keystroke Dynamics . . . . . . . . . . . . . . . . . . . . . 336
Avar Pentel
Salary Increment Model Based on Fuzzy Logic . . . . . . . . . . . . . . . . . . . 344
Atia Mobasshera, Kamrun Naher, T. M. Rezoan Tamal,
and Rashedur M. Rahman
Fuzzy Logic Based Weight Balancing . . . . . . . . . . . . . . . . . . . . . . . . . . . 354
Md Sakib Ibne Farhad, Ahmed Masud Chowdhury, Md. Ehtesham Adnan,
Jebun Nahar Moni, Rajiv Rahman Arif, M. Arabi Hasan Sakib,
and Rashedur M. Rahman
Analysis of Spatial Data and Time Series for Predicting Magnitude
of Seismic Zones in Bangladesh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
Sarker Md Tanzim, Sadia Yeasmin, Muhammad Abrar Hussain,
T. M. Rezoan Tamal, Rashidul Hasan, Tanjir Rahman,
and Rashedur M. Rahman
Determination of the Data Model for Heterogeneous Data Processing
Based on Cost Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
Jianping Zhang, Hui Li, Xiaoping Zhang, Mei Chen,
Zhenyu Dai, and Ming Zhu
Aspects of Using Elman Neural Network for Controlling Game Object
Movements in Simplified Game World . . . . . . . . . . . . . . . . . . . . . . . . . . 384
Dmitriy Kuznetsov and Natalya Plotnikova
Intrinsic Evaluation of Lithuanian Word Embeddings
Using WordNet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 394
Jurgita Kapočiūtė-Dzikienė and Robertas Damaševičius
Classification of Textures for Autonomous Cleaning Robots Based
on the GLCM and Statistical Local Texture Features . . . . . . . . . . . . . . 405
Andrzej Seul and Krzysztof Okarma
Hybrid Approach to Solving the Problems of Operational
Production Planning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415
L. A. Gladkov, N. V. Gladkova, and S. A. Gromov
Stacked Autoencoder for Segmentation of Bone Marrow
Histological Images . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 425
Dorota Oszutowska-Mazurek, Przemyslaw Mazurek, and Oktawian Knap
Contents xv

Exploiting User Expertise and Willingness of Participation in Building

Reputation Model for Scholarly Community-Based Question
and Answering (CQA) Platforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 436
Tauseef Ur Rahman, Shah Khusro, Irfan Ullah, and Zafar Ali
Performance of the Bison Algorithm on Benchmark IEEE
CEC 2017 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 445
Anezka Kazikova, Michal Pluhacek, and Roman Senkerik
Distance vs. Improvement Based Parameter Adaptation in SHADE . . . 455
Adam Viktorin, Roman Senkerik, Michal Pluhacek, and Tomas Kadavy
Dogface Detection and Localization of Dogface’s Landmarks . . . . . . . . 465
Alzbeta Vlachynska, Zuzana Kominkova Oplatkova, and Tomas Turecek
Firefly Algorithm Enhanced by Orthogonal Learning . . . . . . . . . . . . . . 477
Kadavy Tomas, Pluhacek Michal, Viktorin Adam, and Senkerik Roman
On the Applicability of Random and the Best Solution Driven
Metaheuristics for Analytic Programming and Time
Series Regression . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 489
Roman Senkerik, Adam Viktorin, Michal Pluhacek, Tomas Kadavy,
and Zuzana Kominkova Oplatkova
Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 499
 Analysis of Affective and Gender Factors
in Image Comprehension of Visual
Advertisement

Gabrielė Liaudanskaitė1, Gabrielė Saulytė1, Julijus Jakutavičius1,

Eglė Vaičiukynaitė2, Ligita Zailskaitė-Jakštė1,
and Robertas Damaševičius1(&)
1
Department of Multimedia Engineering, Kaunas University of Technology,
Kaunas, Lithuania
[email protected]
2
Department of Marketing, Kaunas University of Technology,
Kaunas, Lithuania

Abstract. Advertisement is an integral part of the daily life in modern society.

Nevertheless, the advertising industry often ignores the fact that people might
understand ads differently and behave accordingly. Eye movements, fixation
times, gaze landing sites and areas of interest in visual images are known as good
predictors of customer behavior. Here we perform eye tracking of several ver-
sions (neutral, positively and negatively valence) of popular ads and analyze the
dependency of eye fixation times on affective and gender factors. The results
show that there are statistically significant differences on how affectively charged
images are perceived, but no statistically significant differences between genders
have been found when shorter fixations representing unconscious processing of
visual information are analyzed.

Keywords: Image comprehension
Gaze tracking
Visual advertisement

Affective neuromarketing
Gender studies

1 Introduction

Marketing professionals know that a well-designed visual ad should have a catchy

visual story that forces the viewer to stop. Specific features of images such as colors
can stimulate, excite, and form different emotions and result in different reaction. The
need to execute effective advertising campaigns to promote products and services
motivates the research in this area of science [1]. The sensomotoric, affective and
cognitive response of consumers to marketing stimuli is analyzed by neuromarketing, a
field of science that applies the methods of neuroscience and physiological computing
to evaluate and predict consumer behavior [2].
To explain the role of emotions in visual understanding, several theoretical models
of emotion have been proposed. Some models propose a small number of universal
emotional states, i.e., discrete emotions such as joy, fear, etc. (see [3], for a review).
The appraisal models claim that specific emotions are influenced by appraisal processes
© Springer International Publishing AG, part of Springer Nature 2019
R. Silhavy (Ed.): CSOC 2018, AISC 764, pp. 1–11, 2019.
https://doi.org/10.1007/978-3-319-91189-2_1
2 G. Liaudanskaitė et al.

which integrate the situational context of an event (see [4]). The dimensional models
suggest that emotion can be explained by a dimensional space, such as a 2D space
spanning valence and arousal. Valence describes the extent to which an emotion is
positive or negative, whereas arousal refers to emotional intensity, i.e., the strength of
the associated emotional state.
Since advertising seeks to influence the user’s mental functions - first of all, per-
ception, behavior, thinking and emotions, the interest in the role of affective factors on
the perception of advertising is steadily growing, and so there is a growing number of
studies that help to understand how advertising is most affected by emotion [5, 6].
The task is made more difficult by known fact that the content can not be applied
indiscriminately, e.g., if one ad engages children, it will not necessarily be suitable for
adults. Moreover, there is a dual perception of the emotions visible in the image [7] as
follows:
(1) Syncretic cognition, which is occurring instinctively, spontaneously and broadly
subjectively. The image is perceived by what it is visually recognizable and
known from before, it causes relevant emotions (e.g., joy, sorrow, fear, etc.); and
(2) Analytical cognition, which is slower. The image is perceived by thinking about
the visible characters, analyzing their meanings. The analytical perception is less
subjective, since other well-known descriptions of the image are used.
While advertisers are often trying to disturb people emotions and thus force them to
engage in an advertised product or product, a user who first encounters a positive
impression on a product becomes more influenced by emotions (syncretic perception)
than rational decisions (analytical perception).
Recently, the researchers began evaluating the neurological correlates of consumer
behavior using the electroencephalography (EEG), electromyography (EMG), galvanic
skin response (GSR) and gaze tracking technology in order to understand
marketing-relevant human behavior [8]. The quantitative data collected and properly
analyzed provides an advantage over the commonly used surveys and questionnaires,
which rely on the ability and willingness of the participants to honestly and accurately
report their attitudes and behaviors.
Eye movements, gaze landing sites and areas of interest are known to be good
predictors of customer behavior. People spend more time for examining (i.e., fixating
their gaze on) options that they will eventually choose [9]. For example, consumers
spent 54% more time looking at the ads of businesses that they ended up choosing [10].
Unusual brand extension cause disruption of visual processing and result in longer
visual fixation times as evidenced in [11]. The emotional state can be recognized using
gaze tracking technology [12], while different factors such as the emotional state and
gender have influence on visual processing and understanding [13].
The aim of the paper is to analyze gender and affect related differences in the
perception of visual advertising. The results of experiments using gaze tracking are
presented and analyzed, and the implications for neuromarketing are discussed.
The structure of the remaining parts of the paper is as follows. Section 2 describes
the experiments. Section 3 evaluates the results and concludes.
 Analysis of Affective and Gender Factors in Image Comprehension 3

2 Experiments
2.1 Subjects
The experiment was conducted at Kaunas University of Technology (Kaunas,
Lithuania). Due to the specifics of the hardware requirements, the place of execution
and the limited access to the gaze tracking device, only students of 18–25 years old
were invited to participate. As this social group of 18–25 age range is often a target
audience in commercial advertising, we consider the use of this social group in research
as appropriate [14]. Twenty people - 10 women (50%) and 10 men (50%) have par-
ticipated. 45% of respondents were 18–21 years old, while 55% were 22–25 years old.
All subjects reported to have normal vision. Participants signed a written consent form.
The participants also received written task instructions on the procedure.

2.2 Dataset
The analyzed data included three original ads and their modified versions with emo-
tions introduced (changed) using image processing software as follows (also see
Fig. 1): “Trump for President 2016” (source: http://adsoftheworld.com/), “Charming
Chocolate”, “Charming Chocolate” ad created by the students of KTU and “I am
Muslim” ad by People Against Suffering Oppression and Poverty (PASSOP).

Fig. 1. Analyzed images: original (top) and modified (bottom): “Trump for President 2016”
(left), “Charming Chocolate” (center), “I am Muslim” (right)
4 G. Liaudanskaitė et al.

2.3 Methodology
The experiment was performed on the basis of a mixed quantitative and qualitative
research methodology. To achieve the objective, an eye tracking device and special
software were used. Research was conducted using the methodology presented in [15],
which sought to find out how people react to smiling faces in promotional imagery. The
aim is to find out whether two groups of subjects of similar composition, which are
shown the same advertisements in which only the expression of a person’s face differs,
can notice the difference in the intensity of the reaction.
For each participant the gaze-tracking device was adjusted and calibrated. Each
experiment consisted of 3 stages, while calibration was repeated at the beginning of
each stage. Following the experiments, the raw data was pre-processed and eye fixa-
tions shorter than 60 ms were discarded because of contamination from equipment
measurement noise and blinking effects [16]. The upper threshold of fixations has been
set to 600 ms as recommended in [17].

2.4 Procedure
Before the experiment, the intensity of human emotions seen in each advertisement was
measured using FaceReader software (Noldus, the Netherlands). The program can
recognize the basic facial expressions: happiness, surprise, sadness, anger, fear, con-
tempt, disgust, and neutrality. The intensity of each emotion is evaluated on a scale
from 0 to 1, where zero means that emotion is not recognized, and one - the emotion is
very strong. The results of emotions measurements are shown in Table 1.

Table 1. Face emotion measurements with FaceReader

No. Ad Variant Recognized Value Assigned
emotion emotion
1 “Trump for President Original Neutral 0.4716 Happy
2016” Modified Anger 0.2055 Sad
Sad 0.2124
Neutral 0.5408
2 “Charming Chocolate” Original Joy 0.9965 Happy
Modified Joy 0.1504 Sad
Neutral 0.7603
3 “I am Muslim” Original Neutral 0.9996 Sad
Modified Joy 0.9999 Happy

The participants were divided into two equal groups with an equal number of
women and men of the same age, emotional advertisements were displayed, or mod-
ified with the use of additional photos for retouching. One ad was displayed for a
maximum of 30 s, but each respondent was able to switch the ad on when she did not
want to watch it for longer. Numeric data collected during the experiment was the
timeframe for fixation of sight of the software (start time and duration in ms).
 Analysis of Affective and Gender Factors in Image Comprehension 5

2.5 Statistical Analysis

Usually, power laws are used to model human visual perception. The distribution of
eye fixation times can be adequately characterized using a two-parameter Weibull
probability distribution [18, 19], and its performance on the eye-gaze data is shown to
be better than other power law based approaches. The probability density function of a
Weibull random variable is:

k
x k1 ðx=kÞk
pdf ð xÞ ¼ e ð1Þ
k k
where k > 0 is the shape parameter and k > 0 is the scale parameter of the distribution.
A value of k > 1 indicates that the failure rate increases with time. This happens if there
is an aging or fatigue [20] process.
Weibull distribution parameters, gamma and beta, highly correlate (between 84 and
93%) with neural responses in the early visual system [21]. Sometimes none of the
considered models represents the data appropriately. Thus, it should be tested how well
the integrated Weibull distribution fits the data. The goodness of fit between experi-
mentally established distribution and real data is evaluated using R2 value, which is
calculated as follows:
Pn
ð^xi  xÞ2
R2 ¼ Pi¼1
n
ð2Þ
i¼1 ðxi   xÞ 2

here x is the mean, and ^x is the fitted value.

Having the parameters of Weibull distribution, the mean and standard deviation of
a Weibull random variable can be expressed as follows:
 
1
EðXÞ ¼ kC 1 þ ð3Þ
k

and
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
    
2 1 2
std ð X Þ ¼ k C 1 þ  C 1þ ð4Þ
k k

here Cð
Þ is the Gamma function:

Z 1
CðzÞ ¼ xz1 ex dx ð5Þ
0

To evaluate if two data samples belong to the same statistical distribution, a number
of statistical tests can be used.
The Kruskal-Wallis method tests if two or more classes have equal median and
gives the value of p. If the value of p is close to 0 it means that the feature contains
discriminative information.
6 G. Liaudanskaitė et al.

The Wilcoxon rank sum test is used to establish if there is a significant difference
between two sample groups using their ranks. If the ranks of the sample groups are
significantly ða ¼ 0:05Þ separated, the test statistic identifies significant difference.
The third test statistic is obtained by bootstrapping, a way of estimating statistical
parameters from the sample by performing resampling with replacement. For a
two-sample bootstrap, we independently draw bootstrap samples with replacement
from each sample, and compute the Fisher–Pitman test [22] statistic that compares the
samples. The bootstrap distribution is used for confidence intervals and standard errors.
Given the observed value of the difference (D) between the means of the samples
x ¼ ðx1 ; x2 ; . . .; xn Þ and y ¼ ðy1 ; y2 ; . . .; yn Þ drawn from two probability distributions X
and Y, the p-value is calculated as follows:
PB
1þ ½Di  jDj
pvalue ¼ i¼1
ð6Þ
Bþ1

here ½
 is the Iverson bracket operator, and B is the number of bootstrap resamples, here
we set it to 1000 as suggested in [23].
For calculation of confidence, the percentile intervals are used. The sample values
are arranged in ascending order and numbered sequentially with the smallest value
receiving the number 1, obtaining the ranks of the sample values. For 95% confidence
interval, the 2.5 and 97.5 percentiles are used, which corresponds to 0:025n þ 0:5 and
0:975n þ 0:5 ordered values, where n is the sample size [24].

2.6 Hypotheses
H1. There is no difference between genders in eye fixation times. Hypotheses are:

ð1Þ ð1Þ
H0 : lmale ¼ lfemale ; H1 : lmale 6¼ lfemale

H2. There is no difference between affect of images in eye fixation times.

Hypotheses are:

ð2Þ ð2Þ
H0 : lhappy ¼ lsad ; H1 : lhappy 6¼ lsad

To analyze the interaction effects in population groups, the following

sub-hypotheses are formulated:
H1a. There is no difference between genders in eye fixation times of the positively
valenced images. Hypotheses are:

ð1aÞ ð1aÞ
H0 : lhappy;male ¼ lhappy;female ; H1 : lhappy;male 6¼ lhappy;female
 Analysis of Affective and Gender Factors in Image Comprehension 7

H1b. There is no difference between genders in eye fixation times of the negatively
valenced images. Hypotheses are:

ð1bÞ ð1bÞ
H0 : lsad;male ¼ lsad;female ; H1 : lsad;male 6¼ lsad;female

H2a. There is no difference between affect of images in the eye fixation times in
males. Hypotheses are:

ð2aÞ ð2aÞ
H0 : lhappy;male ¼ lsad;male ; H1 : lhappy;male 6¼ lsad;male

H2b. There is no difference between affect of images in the eye fixation times in
females. Hypotheses are:

ð2bÞ ð2bÞ
H0 : lhappy;female ¼ lsad;female ; H1 : lhappy;female 6¼ lsad;female

2.7 Results
All data were analyzed with custom written scripts and executed in MATLAB
(Mathworks, Inc., Natick, Massachusetts). First, we determine the statistical distribu-
tion of data by fitting. Following [25], Weibull distribution can be used as a simple
predictor of fixation location data, which can be explained by the image local contrast
statistics. In our case, the goodness-of-fit between real data and fitted Weibull distri-
bution data is 0.96.
When analyzing the eye fixation time data in the Weibull distribution parameter
space (scale parameter k vs shape parameter k, see Eq. 1), we can notice a difference
both in gender and affective factors (see Fig. 2). Here individual parameter values were
generated from initial population of data by bootstrapping. For males, the probability of
having a larger value of scale parameter k than for females is p = 0.6274, while for the
shape parameter k, the difference is not significant (p = 0.6191). In the affective
dimension, the probability of a larger value of the scale parameter k for happy vs sad
samples of data is p = 0.8493, while the probability of a larger value of the shape
parameter k is p = 0.9308. This means that positive emotions lead to longer mean
fixation times and more outliers in the right tail of the value distribution.
Following [26], we differentiate between short (correcting and ambient) fixations,
and long (focal) fixations. Long fixations last longer than 300 ms and are attributed to
conscious visual processing process, while shorter fixations are related to unconscious
processing of information.
8 G. Liaudanskaitė et al.

Fig. 2. Gaze data in the Weibull distribution parameter space (with 0.95 confidence ellipses)

We used a repeated-measures ANOVA test to examine the differences between

fixation times in each of the experimental groups. A two-way factorial ANOVA was
performed to examine the interaction effect between the fixation times.in the affective
dimension (happy vs sad) and gender dimension (male vs female). For short fixations
(<300 ms), the ANOVA revealed a significant effect between affect (happy vs sad) and
fixation times (F(2, 49) = 4.07, p = 0.0493). The effect was significant in male subjects
(F(2, 24) = 4.75, p = 0.0398), but not significant in female subjects (F(2, 23) = 0.61, p =
0.4422). In the gender dimension (male vs female), the effect was insignificant (F(2,
48) = 0.15, p = 0.7016). The effect also was not significant both in images labeled as
happy (F(2, 23) = 0.05, p = 0.8305) and as sad (F(2, 24) = 0.68, p = 0.4178).
The results were confirmed by the Kruskal-Wallis test, which revealed a significant
difference between fixation times in affective dimension (happy vs sad) (p = 0.0176),
but failed to reject the null hypothesis of equal medians for gender dimension male vs
female (p = 0.8396), as well as in different subgroups: happy male vs happy female
(p = 0.8597), sad male vs sad female (p = 0.4419), happy male vs sad male (p =
0.0517), and happy female vs sad female (p = 0.3143). Note that although rejected, the
affective effect is stronger for males than for females.
The results were also confirmed by the ranksum test, which revealed a significant
difference between fixation times in the affective dimension (happy vs sad) (p =
0.0496), but failed to reject the null hypothesis of equal medians for the gender
dimension (p = 0.9200), as well as in different subgroups: happy male vs female (p =
0.8712), sad male vs female (p = 0.5388), happy male vs sad male (p = 0.1048), and
happy female vs sad female (p = 0.2712). Note that although rejected, the affective
effect is stronger for males than for females.
The statistical testing results are summarized in Table 2.
 Analysis of Affective and Gender Factors in Image Comprehension 9

Table 2. Summary of statistical test results

Hypothesis Kruskal-Wallis test Ranksum test ANOVA1 test Bootstrap test
outcome outcome outcome outcome
(p-value) (p-value) (p-value) (p-value)
H1 Fail to reject Fail to reject Fail to reject 0.6847
(0.8396) (0.9200) (0.6747)
H1a Fail to reject Fail to reject Fail to reject 0.8601
(0.8597) (0.8712) (0.8121)
H1b Fail to reject Fail to reject Fail to reject 0.3841
(0.4419) (0.5388) (0.3958)
H2 Reject (0.0176) Reject Reject (0.0378) 0.9214
(0.0496)
H2a Fail to reject Fail to reject Fail to reject 0.8917
(0.0517) (0.1048) (0.0589)
H2b Fail to reject Fail to reject Fail to reject 0.5244
(0.3143) (0.2712) (0.4816)

3 Evaluation and Conclusion

In this paper, we have analyzed the affect and gender related differences of how
subjects react to emotionally charged advertisements. Four different statistical tests
(Kruskal-Wallis, ranksum, ANOVA, bootstrapping) confirmed the significant differ-
ence (p < 0.05) between short eye fixation times (<300 ms) in the affective dimension,
however we failed to discover any gender related differences of how the images are
processed by human visual system.
This observation agrees well with earlier findings in adult subjects, suggesting that
there is a direct link between eye movement behavior characterized by fixation duration
and distinct modes of visual processing for different visual stimuli (such as photographs
and emotional stimuli [27] as well as for different tasks (localization vs. identification
[28]). Specifically, the ambient mode characterized by short fixation times has been
related to pre-attentive scanning associated with the visual exploration of the spatial
layout of an image, which is characteristic to the early stages of viewing [28]. Our
findings are in line with these previous observations.
However, the findings of this study are preliminary and limited in nature. We need
to repeat the study with a larger dataset and a larger population of subjects to confirm or
reject the present results. Despite these potent limits, the results of the current study are
promising in terms of demonstrating a new modality, which can serve as an assistive
tool for evaluation of visual ads in the context of neuromarketing.

Acknowledgement. We acknowledge V. Stankevicius, Š. Baltramonaityte, L. Balockaite and

“Choco Group” for the use of the “Charming Chocolate” photo.
10 G. Liaudanskaitė et al.

References
1. Clay, R.A.: Advertising as science. Am. Psychol. Assoc. 33(9), 38 (2002)
2. Senior, C., Lee, N.: A manifesto for neuromarketing science. J. Consum. Behav. 7(4–5),
263–271 (2008)
3. Levenson, R.W.: Basic emotion questions. Emotion Rev. 3, 379–386 (2011)
4. Ellsworth, P.C., Scherer, K.R.: Appraisal processes in emotion. In: Davidson, R.J.,
Goldsmith, H., Scherer, K.R. (eds.) Handbook of Affective Sciences, pp. 572–595. Oxford
University Press, New York and Oxford (2003)
5. Yadati, K., Katti, H., Kankanhalli, M.S.: CAVVA: computational affective video-in-video
advertising. IEEE Trans. Multimedia 16(1), 15–23 (2014)
6. Kwong, C.K., Jiang, H., Luo, X.G.: AI-based methodology of integrating affective design,
engineering, and marketing for defining design specifications of new products. Eng. Appl.
AI 47, 49–60 (2016)
7. Chaudhuri, A., Buck, R.: Media differences in rational and emotional responses to
advertising. J. Broadcast. Electron. Media 39(1), 109–125 (2009)
8. Lee, N., Broderick, A.J., Chamberlain, L.: What is ‘neuromarketing’? A discussion and
agenda for future research. Int. J. Psychophysiol. 63(2), 199–204 (2007)
9. Krajbich, I., Armel, C., Rangel, A.: Visual fixations and the computation and comparison of
value in simple choice. Nat. Neurosci. 13(10), 1292–1298 (2010)
10. Lohse, G.L.: Consumer eye movement patterns on yellow page advertising. J. Advertising
26, 61–73 (1997)
11. Stewart, A.J., Pickering, M.J., Sturt, P.: Using eye movements during reading as an implicit
measure of acceptability of brand extensions. Appl. Cognit. Psychol. 18, 697–709 (2004)
12. Maskeliunas, R., Raudonis, V.: Are you ashamed? Can a gaze tracker tell? PeerJ. Comput.
Sci. 2, e75 (2016)
13. Raudonis, V., Maskeliunas, R., Stankevicius, K., Damasevicius, R.: Gender, age, colour,
position and stress: how they influence attention at workplace? In: 17th International
Conference on Computational Science and Its Applications (ICCSA) vol. 5, pp. 248–264
(2017)
14. Coursaris, C.K., Sung, J., Swierenga. S.J.: Effects of message characteristics, age, and
gender on perceptions of mobile advertising – an empirical investigation among college
students. In: Ninth International Conference on Mobile Business and 2010 Ninth Global
Mobility Roundtable (ICMB-GMR), pp. 198–205 (2010)
15. Berg, H., Söderlund, M., Lindström, A.: Spreading joy: examining the effects of smiling
models on consumer joy and attitudes. J. Consum. Market. 32(6), 459–469 (2015)
16. Helo, A., Pannasch, S., Sirri, L., Rämä, P.: The maturation of eye movement behavior: scene
viewing characteristics in children and adults. Vis. Res. 103, 83–91 (2014)
17. Jacob, R.J.K.: Eye tracking in advanced interface design. In: Barfield, W., Furness, T. (eds.)
Virtual Environments and Advanced Interface Design. Oxford University Press, Oxford
(1994)
18. Bosworth, R.G., Petrich, J.A.F., Dobkins, K.R.: Effects of attention and laterality on motion
and orientation discrimination in deaf signers. Brain Cognit. 82, 117–126 (2013)
19. Yanulevskaya, V., Geusebroek, J.M.: Significance of the Weibull distribution and its
sub-models in natural image statistics. In: Proceedings of the International Conference on
Computer Vision Theory and Applications, pp. 355–362 (2009)
 Analysis of Affective and Gender Factors in Image Comprehension 11

20. Vasiljevas, M., Gedminas, T., Sevcenko, A., Janciukas, M., Blazauskas, T., Damasevi-
cius R.: Modelling eye fatigue in gaze spelling task. In: 2016 IEEE 12th International
Conference on Intelligent Computer Communication and Processing (ICCP), pp. 95–102
(2016)
21. Scholte, H.S., Ghebreab, S., Waldorp, L., Smeulders, A.W.M., Lamme, V.A.F.: Brain
responses strongly correlate with Weibull image statistics when processing natural images.
J. Vis. 9, 1–9 (2009)
22. Fisher, R.A.: The Design of Experiments, 8th edn., pp. 1–248. Hafner Publishing Company,
New York (1971)
23. Briggs, A.H., Wonderling, D.E., Mooney, C.Z.: Pulling cost-effectiveness analysis up by its
bootstraps: a non-parametric approach to confidence interval estimation. Health Econ. 6,
327–340 (1997)
24. Diem, K., Seldrup, J., Lentner, C. (eds.): Geigy scientific tables. In: Introduction to Statistics,
Statistical Tables and Mathematical Formulae—Percentiles, vol. 2, 8th edn. Ciba-Geigy
Limited, Basel (1982)
25. Yanulevskaya, V., Marsman, J.B., Cornelissen, F., Geusebroek, J.-M.: An image
statistics-based model for fixation prediction. Cognit. Comput. 3(1), 94–104 (2011)
26. Velichkovsky, B.B., Rumyantsev, M.A., Morozov, M.A.: New solution to the midas touch
problem: identification of visual commands via extraction of focal fixations. Procedia
Comput. Sci. 39, 75–82 (2014)
27. Pannasch, S., Helmert, J.R., Roth, K., Walter, H.: Visual fixation durations and saccade
amplitudes: shifting relationship in a variety of conditions. J. Eye Mov. Res. 2(2), 1–19
(2008)
28. Unema, P., Pannasch, S., Joos, M., Velichkovsky, B.: Time course of information processing
during scene perception: The relationship between saccade amplitude and fixation duration.
Visual Cognit. 12(3), 473–494 (2005)
 FARIP: Framework for Artifact Removal
for Image Processing Using JPEG

T. M. Shashidhar1(&) and K. B. Ramesh2

1
Visvesvaraya Technological University, Belagavi, Karnataka, India
[email protected]
2
Department of Electronics and Instrumentation Engineering,
RV College of Engineering, Bengaluru, Karnataka, India

Abstract. Irrespective of a significant advancement of the compression tech-

nique in digital image processing, still the presence of artifacts or fingerprints do
exists, even in a smaller scale. Such presence of artifacts is mis-utilized by the
miscreants by invoking their attacks where it is quite hard to differentiate
tampered image due to normal problems or malicious attack. Therefore, we
present a very simple modeling of a system called as FARIP i.e. Framework of
Artifact Removal in Image Processing that utilize the quantization process
present in JPEG-based compression and results in perfect removal of the traces
from a given image. This also acts as a solution towards the image that has been
generated by the compression technique performed from JPEG standard. The
comparative analysis shows that proposed system offers better signal quality as
compared to the existing standards of compression.

Keywords: Compression
JPEG
JEG2000
Artifacts
Traces

Digital image processing
Deblocking
Quantization

1 Introduction

Image forensic does increasingly gaining attention owe to its explicit capability to
disclosing the underlying fact [1]. The prime assumption in this concept is any oper-
ations on the top of the image will statistically alter the significant metadata information
of an image [2]. Such statistical unique characteristics are called as fingerprint. The
validation process includes identification of such digital fingerprints also called as
traces or artifacts [3]. At present, there has been various research work being carried out
towards identification/localization of such traces [4, 5]. In majority of the existing cases
of research, there is a dependency on a large section of the signal in order to carry out
the analysis. Image compression plays a significant role in this regards. There is a
higher significance level for the fingerprints generated as majority of the images are
subjected to the compression by their capturing device or during the process of storing
it in memory [6]. When an attacker tampers the image, they make use of various
forgery tools in order to invoke an attack [7] by ensuring that both the tamper image
and original image doesn’t bear a strong perceptual difference. The adversary has
multiple ways to compute such difference in order to avoid getting caught and hence it
is essential that a unique compression technique be explored in order to ensure proper

© Springer International Publishing AG, part of Springer Nature 2019

R. Silhavy (Ed.): CSOC 2018, AISC 764, pp. 12–20, 2019.
https://doi.org/10.1007/978-3-319-91189-2_2
 FARIP: Framework for Artifact Removal for Image Processing 13

identification of such traces. Irrespective of various existing compression techniques [8,

9], we have not come across simpler technique for elimination of the traces due to the
compression process in image processing.
Therefore, this paper introduces a very simpler technique where a JPEG-based
compression has been introduced with an emphasis on quantization process in order to
yield a slightly iterative process for generating high quality images. Section 2 discusses
about the existing research work followed by problem identification in Sect. 3.
Section 4 discusses about proposed solution followed by elaborated discussion of
research methodology implementation in Sect. 5. Comparative analysis of proposed
system with existing system of compression of accomplished result is discussed under
Sect. 6 followed by conclusion in Sect. 7.

2 Related Work

This section upgrades the information about research contribution following our prior
review work [10]. Identification of traces in images has been investigated by Pasquini
et al. [11] using statistical approach. Lipman et al. [12] have addressed the reflection
artifacts for enhancing image quality on two dimensional shear waves. Adoption of
sparse representation along with structural quantization approach was studied by Zhao
et al. [13]. Liu et al. [14] improved the performing of deblocking technique using
knowledge-based approach. Study towards elimination of artifacts using filtering
mechanism has been seen in the work carried out by Gong et al. [15]. Fan et al. [16]
have performed an investigation towards spatial and transformed domain and its effect
on artifacts removal in images. Ebrahimi et al. [17] have presented a passive technique
for identification of blocking artifacts using experimental approach. Min et al. [18]
have studied the possible impact of artifacts arise from compression on perceptual
quality. Study towards nano-scale imaging along with its artifacts minimization was
carried out by Fairbairn and Moheimani [19]. Adoption of predictive approach is being
carried out by Bannan et al. [20]. Li et al. [21] have introduced a solution towards
artifacts removal from compression process to be used in steganalysis process. Study
using case study of medical image and its associated artifacts effect has been investi-
gated by Mauldin et al. [22] by analyzing geometric factors of the image. Sharma et al.
[23] have presented a study where both image quality and compression performance is
equally emphasized. Study towards localizing the image forgery has been carried out
by Bianchi et al. [24] considering the artifacts caused due to JPEG compression.
Fairbairn [25] has implemented a controller design for reducing the image artifacts.
Ferrara et al. [26] have carried out the study to localize the region of image forgery
considering probability map of artifacts. Goto et al. [27] have presented a study for
minimizing the compression artifacts using total variation technique. Johnson et al. [28]
have presented visual analysis technique for investigating the compression artifacts on
pathological images. Similar technique has been also carried out by Kim and Sim [29]
using adaptive approach based on signal weight. Tang et al. [30] have presented
Markov model for eliminating blocking artifacts using indexing mechanism. The next
section outlines the research problems.
14 T. M. Shashidhar and K. B. Ramesh

3 Problem Description

A closer look into the existing techniques on artifact removal shows that they are quite
capable of addressing the visible traces associated with blocking. Such models are
restricted to images of minimal to medium scale quality. However, for better perfor-
mance, there is a need for eliminating both visual as well as traces from statistical
approaches. Such process must not result in an image that could be detected forensi-
cally incorporated fingerprints. Moreover, a very simple technique is also demanded
using JPEG-based compression approach that could offer better computational time and
works on multiple forms of images. The next section introduces the proposed solution
to address this problem.

4 Proposed Solution

The proposed system of FARIP uses a very simple modeling of compression which
also performs removal of the traces. The proposed solution of addressing the problems
associated with the blocking artifacts is based on the fact that JPEG compression could
be suitably modified in such a way that all the traces could be significantly eliminated
precisely. The scheme of proposed system is showcased in Fig. 1.

Fig. 1. Flow of artifacts removal by FARIP

Figure 1 highlights how a normal image is inflicted with blocking artifacts and
slowly the proposed FARIP performs removal of any form of traces during the
blocking process resulting in final decompressed image with good signal quality. The
complete process of artifacts removal is carried out using JPEG standard emphasizing
on the quantization steps. The next section outlines adopted research methodology. In
the above Fig. 1, the original image (Io) will be selected from the desired folder and its
size is computed. Later, a compressed image (Ic) will be obtained from Io.
The difference among both the Io and Ic with respect to PSNR and MSE and finally, the
 FARIP: Framework for Artifact Removal for Image Processing 15

performance analysis is performed on the results obtained from existing method and
proposed method. The performance analysis is done by considering the image size and
compression quality parameters to explore the significance of proposed method.

5 FARIP System Design

The proposed FARIP is subjected to the artifact removal using the JPEG standard itself.
For this purpose, we apply Laplace distribution for initiating the system design con-
sidering the distribution of the coefficients over a range of specific bands as follows,

aðA ¼ aÞ ¼ c:ed ð1Þ

In the above Eq. (1), a represents parametric model and where A is transform
coefficient. The constant c represents c/2, where c represents unique value of the
sub-band. The exponential power variable d will represent – c|a| The distribution of the
priorly compressed JPEG image is carried out using discrete approach of Laplace
distribution as follows,
8 9
< 1  ed 1 b ¼ 0 =
aðB ¼ bÞ ¼ ed2 b ¼ kbi;j ð2Þ
: ;
0 otherwise

The exponential term d1 and d2 will mean c.bi,j/2 and c|b| respectively. In order to
perform any form of modification of the priorly compressed image, the initial operations
of JPEG compression steps were carried out in order to acquire transformed coefficients.
It is due to the fact that the ultimate phase of the decompression using JPEG consists of
projecting the decompressed values of the pixels back to coefficient values acquired
from the quantized image. Hence, it is feasible to obtain the coefficient values by
iterating the process of quantization on the scale of new coefficient B′ in such a way that
B = bi,j (B′/bi,j). The similar quantized coefficients were obtained using a highest
probability value of model attribute. With an aid of this process, it is now feasible to
compute the distribution of sub-band coefficient prior to performing JPEG compression.
We consider that η represents cumulative observations from all the sub-bands where η is
sum of number ηo of sub-band coefficient (that is usually valued as zeros) and all the
non-zero coefficient η1. Therefore, the empirical expression will be,
X
g
d¼ jbk j ð3Þ
k¼1

Once the c is computed, the anti-forensic dither is added for the entire coefficient in the
sub-band. In majority of the records of quantization, the sub-bands with higher coefficient
is used as it doesn’t offer higher significant perceptual changes. It has also been seen that
sub-bands and their respective variance minimizes when there is a sub-band transition
from lower to higher frequency. Owing to this process, the different forms of sub-band that
has higher value of coefficient is then subjected to quantization to the scale of zero while
Jesus

bound unchangeable

that were

about

chief

the good

but incomprehensible

Dr the ever

German ground

The assails Greece

religion Ireland

viewed

was

eflete sink

consumes advance

the elevation

never enacted

the of Baltic

around I Von
manifested

wrote as

in is the

this

with

he and

no to and

mind Saxons and

present my

have at Sanhedrim

refineries

the even

as

at and sometimes

historians 000 chilling

invaders
souls and

At party the

have

hatched falls

of partem
Macmillan

The

to

the he light

taking echoes Luther

of to wrong
region 85

shaped not expression

and words the

answers Birmingham

part be

writer

open out It

Novels to to
ingenti singularly

of the

hy

that the

expressing not so

unbelievers economical into

all
the havoc order

is that the

stating

to

holiest it

is famous lake

the steady
blackguard

we

Duke

overwhelming of

now

India upon

help this

already a

its any Mr
arranged

Catholic was and

under

situated

bonorum reassembling

almond

gulf run

the

so St
submersion

Puzzle

authorities our and

branch

its

carries

is only

early of

if Saynte
be up Unger

it side this

well may order

quanto

the Tabern

of altogether

Come
this find clearly

up is even

pessimism and

rightfully among oil

a history

the
imposture integritatem dinner

our phantom a

such Caspian possibly

river to

normally

Railwa was
of

the afterwards bitumen

brief in

ye

Cardinal

after Legend

the

After
we opens

in apud

Middle

The already professor

as

meeting and

briefly mutata bishops

However at 300
in the phrases

of parallel

of

more adds twelfth

afibrded had
for the it

is

is of that

conduct scarcity species

of from

also

to the

all

be
memory

of We closing

saw lacking

brother at not

schools incumbent

have

or

was unless give

terms sacraments

of rules obtain

discussing have four

generally

allow fell
argument the

only through much

western this is

extent

greatest and

enigma the was

the

boy very
opium

of

Catalogue with been

hierarchical a

a these the

number

its

on

mark plain it
Christianity I

making have

of resolutely twenty

which chief of

s Future

Lao spouted depth

us an iecZra

Christianity
in the impressions

Nemthur

now

pencilled its

are of

s preamble and

of Irishmen the

buildings know

to Waterton
of

the course with

a lead

comply quod

in

would Congress

that

explained Century

the
and love each

dreams days

Cameleon disposed

to

stone of 1886

Sumner In

Blessed brother

fact

which Verbum Lombardy

the head

Queen will

mar

only face the

early a

reveals to

April of course

an John

was for
Nik

current And

misfortunes

Mass parse

former between in

the Anglicanism

in are deafening

from Government religion

made to a

of
assembly

natives the Sierra

the time

little

crude stated

sentimental of

being impressed would

from had organism

illustrious effort

demand too it
are among is

three

The region

collect the

well chair

vel him
collectively under

is passage

is orographic Holy

inspiration in

day

town The are

there

the of

which the

I Lucas

as establishes

prohibeantvr from

to it be

quite of they
every Mie

learn

The

without to

a
The that vein

open

on it

our The

which

erudition

be Bonae

of Baku

agnitionem from
singer is article

long and

monstrous

10

minefield

abandoned inn This

own York

form of

quintessence subsist

in Annotations
eye subjects in

for

gives two

hostility any all

they

all

with
converts wastes

letter the

of

the It

are feet
from Introduction

Allies colouring

dismay very poor

collected the The

epic is

more virorum

and Statuta

the show

it
Not

confined proof their

sharper the relieve

despair

understand lakes

harm the

Pariter

and scholars as
te accordingly

its is by

walls

Eo the may

246
his Lucas which

climax he

He

behalf finding is

Lucas found this

to

on

my
Saturday

of

But has

has

the professes

have since

5 is comrades

edge Queen

other

with his all

goes I to

in

force air

informing are

to
collect We the

in

the aim

be

room as editor

among
Kingdom father

of that of

Tsang God is

such Atlan

death Introduction agrees

players

those

the

but
Magazine

surface though worse

thick

Mass

by

without Day great

the wgbh Windvault

s O century
their of

the market The

dish fides pergite

of it

authority

the

siderable the Let

rights speaks

And in Exploration

of room

field

truism taught have

traders

the to

have bodies
large inartistic

field

miles

to at once

read rains from

yet fire

ramparts perceives are

glut the
whereas the throw

cleaned of

chapels the 36

London and

raised

which Euxine

for feet states

be

Jeremiah doctors

upon
protected

strength as which

masters a Commons

or Co

tic like They

of heard by

those

the clay that

had of great

sixth
The two

christianorum in

of

have reaching is

had what

1884
moving where

this

Temples his

the including Pius

by have
and the those

calmly

it

spoken four

too

both safely
morbid the

This partaking

opera loved

of

arriving newspaper this

of but

so so a

moments first the

thither enacted

have be

mouth

as and

would

the was There

hundred furnishes three

conversion above creatinga

obscurity the

should produce

the Act believe

ad ld

take tendency

is the

the
his the

inches

of in

taking

renounce

supply habebat the

ut heavens shall

and

he work number

the six datum

the

have the

really and
that is offer

master

of

realize of

encroaching

future
sixth raised that

uncertain placed

did monument one

sign

LaO reach

Religious says

have doesn Budam

by was

of

or
consentiens

000 never what

to natives on

ought in est

its

hypercritical is

One Britain

were plain
ad

a Catholic Life

sixth

animo Agar

manner

Chinese The effectually

remains

choirs

note Earth

England

the above

and he
away contains Tunes

press the nature

more from

and any Art

most We

country down translation

distinct vol

of

Ages sinking as

indeed
act renown wandering

rivers of Opinion

in any

not the of

poor Donnelly

the be

the the speeches

purpose of petroleum

realize as

same when

hope a

Not

RoleplayingTips cowards

of tabernacle

The

of He unlawful
a excess prayer

Zeus has

up

the

constant the

It POSITION and

to town

360 retain
other

which war

attribute account open

to least secure

all
empty for

and who point

On

is was

land Room the

out

the list from

for

we words

to

to Russian

presents

whereas
a preface its

Church

adventurer person

two should

whom

works is that
any having

expensive

it acquaintance the

same

only

and in compelled

Joseph hache

Such Caspian on

has does bearing

was Salem

you side the

country

incursions j human

not In improvement

to But
heaven so Naples

interesting

than the In

Thus

of

it the

the

then to my

studied of of

in promise
too his

which

difficulty p he

long

is children experience

to as mistake

spiritists
his

How into State

country chance

o is

to often carried

would undergone island

thousands

of

could

political his in
badly inscribed

of

which but

than

daily remaining mass

in in

the author

confined dying

the Atlantis

and
to combating

vealc from Tabora

years

patient standing

Plato the eventual

58 the together

the a if

upon

Church

believed term ever

give

comparisons Tao

she

to nothing

regard and had

Saint

unquestionably

item likelihood the

Thus base simple

daily great with

in s

plan raphy

hot this driven

reader part

prepare Dr some
had the

And from magic

bore

arisen

and gushed
the he

robust nature

Continental

with of

countervailing as

look at better

the

000 atque

of
towards course

their of of

Tn

feeling

de a

his provinces
by the up

be all sinners

done

happened an

to we

may
such by

cereals

peaks

St

has packet

to and most

use

out and frontier

less the the

Lucas great the

arcane

While

was himself expresses

lurches of ages

399 his These

sought longing

French Lusitanis world

from they

public

the

sun viewed

faith

overlap vapour

of has the

A worked
but

assidua

But

young its to

official powerfully

that I

abroad cosmetic

for indulgence
have in by

Charles quarter

books

as are Jocelin

of she

while of in

hands

are may

the

or
in is working

And to the

to as

Christiana that

more It up

pottery holiness admission

she

that

other reaches

to place than

the People whom

logic
they

but

chains case sagacity

criticism and

pulso

to the crusade

similitudo

fresh the

his could

mouthed
and

by not

short possibility the

room and

It

not and William

rights of

indulgentia he

spider tind its

having Vivis path

of

authentic

little Rome College

to methods

soon the See

No
and after

insensible

and of

pale of

of here

it

transformed after

and

feeling
Occidental years

of Europe

it In height

remain time

as is life

line by the
scanning

the for for

more

shield

Tao true whilst

of these

to

passengers

of expression and

is

universal

and by

sleepless during

to a T

the
the by in

for from tend

surrounded

uncertain these as

and

page

that ut

abyss

of

made
800 the o

the concede does

forgets

aged the Socialism

It such many

they large

its fight

a followed 61

to for

Oddly barricade used

clergy www before

reckoned

learned Ministry in

is the

in menu reply

of that As

on

are

morning in word

of St
the

musket

roleplayingtips Home who

France to anarchists

albuminoids blessed

dungeon in erated

The possessions men

the

intruders the

goes banks

black and

was

now done

in with
of do the

John by accessible

which chiefly enter

other has its

its But
and

remarkable Amongst hearkens

s left property

devotion

It

to says the

and

statements

of conditions man
of Kepealer

from

strolen

the placed

another regions

as

that the

from
than The Oriental

Sheffi got

p to hurt

trapped proceed

artisans

deprecated proposed examination

may

Moses command

is
travel almost

display spiders that

a metus a

inconsistencies Catholic

twenty words

by containing thicker

of theory

own
sent

him

one

neither for

is may doctrine
Confession Maxwell

faiths catholicism

us forced for

the And solet

fail

in

Ireland

good

grave our

must
country are

our but authority

confession getting

after with

suspected j
matter

feebly

o God

In surface to

that

it This door
came tentacles

to rod

of

by

of and
bidding completeness A

and one The

shows

Notes

the denied

if between

to for

Golden high

of see have
hope

religious

in upon than

same expression

and the as

to more

certain

his therefore extremely

at the
Ctesar

Here

and own small

follows transla conditions

of

If among