Computational Intelligence for Pattern Recognition

431 Pages·2018·17.254 MB·English

by  Witold Pedrycz, Shyi-Ming Chen

Studies in Computational Intelligence 777
Witold Pedrycz · Shyi-Ming Chen
Editors
Computational
Intelligence
for Pattern
Recognition
Studies in Computational Intelligence
Volume 777
Series editor
Janusz Kacprzyk, Polish Academy of Sciences, Warsaw, Poland
e-mail: [e-mail]
The series “Studies in Computational Intelligence” (SCI) publishes new develop-
mentsandadvancesinthevariousareasofcomputationalintelligence—quicklyand
with a high quality. The intent is to cover the theory, applications, and design
methods of computational intelligence, as embedded in the fields of engineering,
computer science, physics and life sciences, as well as the methodologies behind
them. The series contains monographs, lecture notes and edited volumes in
computational intelligence spanning the areas of neural networks, connectionist
systems, genetic algorithms, evolutionary computation, artificial intelligence,
cellular automata, self-organizing systems, soft computing, fuzzy systems, and
hybrid intelligent systems. Of particular value to both the contributors and the
readership are the short publication timeframe and the world-wide distribution,
which enable both wide and rapid dissemination of research output.
More information about this series at http://www.springer.com/series/7092
Witold Pedrycz Shyi-Ming Chen
(cid:129)
Editors
Computational Intelligence
for Pattern Recognition
123
Editors
Witold Pedrycz Shyi-Ming Chen
University of Alberta National Taiwan University of Science
Edmonton, AB andTechnology
Canada Taipei
Taiwan
ISSN 1860-949X ISSN 1860-9503 (electronic)
Studies in Computational Intelligence
ISBN978-3-319-89628-1 ISBN978-3-319-89629-8 (eBook)
https://doi.org/10.1007/978-3-319-89629-8
LibraryofCongressControlNumber:2018937687
©SpringerInternationalPublishingAG,partofSpringerNature2018
Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpart
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
orinformationstorageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilar
methodologynowknownorhereafterdeveloped.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publicationdoesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfrom
therelevantprotectivelawsandregulationsandthereforefreeforgeneraluse.
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
authorsortheeditorsgiveawarranty,expressorimplied,withrespecttothematerialcontainedhereinor
for any errors or omissions that may have been made. The publisher remains neutral with regard to
jurisdictionalclaimsinpublishedmapsandinstitutionalaffiliations.
Printedonacid-freepaper
ThisSpringerimprintispublishedbytheregisteredcompanySpringerInternationalPublishingAG
partofSpringerNature
Theregisteredcompanyaddressis:Gewerbestrasse11,6330Cham,Switzerland
Preface
Since the inception offuzzy sets, fuzzy pattern recognition, including its method-
ology, algorithms, and applications, has been at the center of the developments
of the technology of fuzzy sets. One can refer here to the seminal paper entitled
Abstraction and pattern classification authored by Bellman, Kalaba, and Zadeh,
whichhasopenedunchartedresearchareasandofferedanewattractiveinsightinto
theprinciplesofpatternclassification.Asofnow,patternrecognitionaugmentedby
the methodology and algorithms of fuzzy sets has established itself as a mature,
well-developedresearchdiscipline with avariety ofadvanced applications.Pattern
recognition comes with a great deal of challenges, exhibits a continuous paradigm
shift (quite often dictated by new applications), and becomes vividly manifested
through a growing diversity of areas of its usage. All of those call for substantial
enhancements of the existing fundamentals or the formation of new paradigms.
Computational Intelligence (CI) with its impressive armamentarium of
methodologiesandtoolsispositionedinauniquewaytoaddressthegrowingneeds
of pattern recognition. As a matter of fact, this can be accomplished in several
tangible ways realized both at the methodological and algorithmic level. There are
atleastfivedominantmanifestationsofCIintherealmofpatternrecognition.They
are associated with: (i) coping with a large volume of data and their diversity,
(ii) setting a suitable level of abstraction, (iii) dealing with a distributed nature of
data along with associated requirements of privacy and security, (iv) building
efficient featurespaces, and (v) building interpretable findings ofclassification ata
suitable level of abstraction.
The key objective of the proposed volume is to provide the community with a
comprehensive and up-to-date treatise in the area of pattern recognition and com-
putational intelligence. It covers a spectrum of methodological and algorithmic
issues, discusses implementations and case studies, identifies the best design
practices, and assesses business models and practices of pattern recognition in
industry, health care, administration, and business. The collection of contributions
forming the edited volume offers the reader a representative view at the progress
and accomplishments of the area with a timely, in-depth, and comprehensive
v
vi Preface
material on the conceptually appealing and practically sound methodology and
practices of CI-based pattern recognition.
ThebookengagesawealthofmethodsofCI,bringsnewconcepts,architectures
andpracticeoffuzzysets,neurocomputing,andbiologicallyinspired optimization.
The chapters cover a wealth of ideas, algorithms, and applications and are a tes-
timonytothesynergisticlinkageswithintheCIareaandCIandpatternrecognition.
Given the leading theme of this undertaking, the book is aimed at a broad
audienceofresearchersandpractitioners.Thankstothenatureofthematerialbeing
coveredandthewaythemainthreadshavebeenorganized,thevolumewillappeal
tothewell-establishedcommunitiesincludingthoseactiveinvariousdisciplinesin
which pattern recognition plays a central role and serves as an efficient vehicle to
produce solutions to numerous classification problems and augments solutions
constructed with the aid of the “standard” methodology and algorithms of pattern
recognition.
With the required prerequisites covered, the book caters to the broad
readership. Those involved in operations research, management, various branches
of engineering, sciences, data science, medicine, and bioinformatics will benefit
from the exposure to the subject matter.
We would like to take this opportunity to express our sincere thanks to the
contributors to the volume for sharing results of their advanced, far-reaching, and
original research, and delivering their views at the rapidly expanding areas of
fundamental and applied research. The reviewers deserve our thanks for their
constructive and timely input. We greatly appreciate a continuous support and
encouragement coming from the Editor-in-Chief, Prof. Janusz Kacprzyk whose
leadership and vision has helped us arrive at the successful completion of this
project.Theeditorial staffatSpringerhasdoneameticulous jobandworkingwith
them was a pleasant experience.
Wehopethatthereaderswillfindthisvolumeinterestingandthevarietyofideas
put forward in this volume will become instrumental in fostering the progress in
research, education, and numerous practical endeavors in the CI-oriented pattern
recognition.
Edmonton, Canada Witold Pedrycz
Taipei, Taiwan Shyi-Ming Chen
Contents
Fuzzy Choquet Integration of Deep Convolutional Neural Networks
for Remote Sensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Derek T. Anderson, Grant J. Scott, Muhammad Aminul Islam,
Bryce Murray and Richard Marcum
Deep Neural Networks for Structured Data. . . . . . . . . . . . . . . . . . . . . . 29
Monica Bianchini, Giovanna Maria Dimitri, Marco Maggini
and Franco Scarselli
Granular Computing Techniques for Bioinformatics Pattern
Recognition Problems in Non-metric Spaces . . . . . . . . . . . . . . . . . . . . 53
Alessio Martino, Alessandro Giuliani and Antonello Rizzi
Multi-classifier-Systems: Architectures,
Algorithms and Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
Peter Bellmann, Patrick Thiam and Friedhelm Schwenker
Learning Label Dependency and Label Preference Relations in
Graded Multi-label Classification. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
Khalil Laghmari, Christophe Marsala and Mohammed Ramdani
Improving Sparse Representation-Based Classification Using Local
Principal Component Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Chelsea Weaver and Naoki Saito
Robust Constrained Concept Factorization . . . . . . . . . . . . . . . . . . . . . . 207
Wei Yan and Bob Zhang
An Automatic Cycling Performance Measurement System Based on
ANFIS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
Andre Vieira Pigatto and Alexandre Balbinot
vii
viii Contents
Fuzzy Classifiers Learned Through SVMs with
Application to Specific Object Detection and Shape Extraction
Using an RGB-D Camera. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253
Chia-Feng Juang and Guo-Cyuan Chen
Particle Swarm Optimization Based HMM Parameter Estimation for
Spectrum Sensing in Cognitive Radio System . . . . . . . . . . . . . . . . . . . . 275
Yogesh Vineetha, E. S. Gopi and Shaik Mahammad
Computational Intelligence for Pattern Recognition in EEG Signals . . . 291
Aunnoy K Mutasim, Rayhan Sardar Tipu, M. Raihanul Bashar,
Md. Kafiul Islam and M. Ashraful Amin
Neural Network Based Physical Disorder Recognition for Elderly
Health Care . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
Sriparna Saha and Raktim Das
Recognizing Subtle Micro-facial Expressions Using Fuzzy Histogram
of Optical Flow Orientations and Feature Selection Methods. . . . . . . . . 341
S. L. Happy and Aurobinda Routray
Improved Deep Neural Network Object Tracking System for
Applications in Home Robotics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369
Berat A. Erol, Abhijit Majumdar, Jonathan Lwowski, Patrick Benavidez,
Paul Rad and Mo Jamshidi
Low Cost Parkinson’s Disease Early Detection and Classification
Based on Voice and Electromyography Signal . . . . . . . . . . . . . . . . . . . . 397
FarikaT.Putri,MochammadAriyanto,WahyuCaesarendra,RifkyIsmail,
Kharisma Agung Pambudi and Elta Diah Pasmanasari
Index .... .... .... .... .... ..... .... .... .... .... .... ..... .... 427
Fuzzy Choquet Integration of Deep
Convolutional Neural Networks
for Remote Sensing
DerekT.Anderson,GrantJ.Scott,MuhammadAminulIslam,
BryceMurrayandRichardMarcum
Abstract Whatdeeplearninglacksatthemomentistheheterogeneousanddynamic
capabilities of the human system. In part, this is because a single architecture is
not currently capable of the level of modeling and representation of the complex
human system. Therefore, a heterogeneous set of pathways from sensory stimu-
lus to cognitive function needs to be developed in a richer computational model.
Herein,weexplorethelearningofmultiplepathways–asdifferentdeepneuralnet-
work architectures–coupled with appropriate data/information fusion. Specifically,
weexploretheadvantageofdata-drivenoptimizationoffusingdifferentdeepnets–
GoogleNet, CaffeNet and ResNet–at a per class (neuron) or shared weight (single
data fusion across classes) fashion. In addition, we explore indices that tell us the
importance of each network, how they interact and what aggregation was learned.
ExperimentsareprovidedinthecontextofremotesensingontheUCMercedand
WHU-RS19datasets.Inparticular,weshowthatfusionisthetopperformer,each
network is needed across the various target classes, and unique aggregations (i.e.,
notcommonoperators)arelearned.
· ·
Keywords Fuzzyintegral Convolutionalneuralnetwork Remotesensing
B
D.T.Anderson( )·G.J.Scott·R.Marcum
ElectricalEngineeringandComputerScience,UniversityofMissouri,Columbia,
MO,USA
e-mail:[e-mail]
G.J.Scott
e-mail:[e-mail]
R.Marcum
e-mail:[e-mail]
M.A.Islam·B.Murray
ElectricalandComputerEngineering,MississippiStateUniversity,
Starkville,MS,USA
e-mail:[e-mail]
B.Murray
e-mail:[e-mail]
©SpringerInternationalPublishingAG,partofSpringerNature2018 1
W.PedryczandS.-M.Chen(eds.),ComputationalIntelligence
forPatternRecognition,StudiesinComputationalIntelligence777,
https://doi.org/10.1007/978-3-319-89629-8_1