ebook img

Smart Computer Vision PDF

358 Pages·2023·11.783 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Smart Computer Vision

EAI/Springer Innovations in Communication and Computing SeriesEditor ImrichChlamtac,EuropeanAllianceforInnovation,Ghent,Belgium The impact of information technologies is creating a new world yet not fully understood. The extent and speed of economic, life style and social changes already perceived in everyday life is hard to estimate without understanding the technological driving forces behind it. This series presents contributed volumes featuring the latest research and development in the various information engi- neering technologies that play a key role in this process. The range of topics, focusing primarily on communications and computing engineering include, but arenotlimitedto,wirelessnetworks;mobilecommunication;designandlearning; gaming;interaction;e-healthandpervasivehealthcare;energymanagement;smart grids;internetofthings;cognitiveradionetworks;computation;cloudcomputing; ubiquitousconnectivity,andinmodegeneralsmartliving,smartcities,Internetof Thingsandmore.Theseriespublishesacombinationofexpandedpapersselected from hosted and sponsored European Alliance for Innovation (EAI) conferences that present cutting edge, global research as well as provide new perspectives on traditional related engineering fields. This content, complemented with open calls forcontributionofbooktitlesandindividualchapters,togethermaintainSpringer’s and EAI’s high standards of academic excellence. The audience for the books consists of researchers, industry professionals, advanced level students as well as practitioners in related fields of activity include information and communication specialists, security experts, economists, urban planners, doctors, and in general representativesinallthosewalksoflifeaffectedadcontributingtotheinformation revolution. Indexing:ThisseriesisindexedinScopus,EiCompendex,andzbMATH. AboutEAI-EAIisagrassrootsmemberorganizationinitiatedthroughcooper- ationbetween businesses,public,privateandgovernment organizations toaddress the global challenges of Europe’s future competitiveness and link the European Research community with its counterparts around the globe. EAI reaches out to hundreds of thousands of individual subscribers on all continents and collaborates with an institutional member base including Fortune 500 companies, government organizations, and educational institutions, provide a free research and innovation platform. Through its open free membership model EAI promotes a new research and innovation culture based on collaboration, connectivity and recognition of excellencebycommunity. B. Vinoth Kumar • P. Sivakumar • B. Surendiran • Junhua Ding Editors Smart Computer Vision Editors B.VinothKumar P.Sivakumar PSGCollegeofTechnology PSGCollegeofTechnology Coimbatore,TamilNadu,India Coimbatore,TamilNadu,India B.Surendiran JunhuaDing Thiruvettakudy UniversityofNorthTexas NationalInstituteofTechnology Denton,TX,USA Puducherry,Karaikal,India ISSN2522-8595 ISSN2522-8609 (electronic) EAI/SpringerInnovationsinCommunicationandComputing ISBN978-3-031-20540-8 ISBN978-3-031-20541-5 (eBook) https://doi.org/10.1007/978-3-031-20541-5 ©TheEditor(s)(ifapplicable)andTheAuthor(s),underexclusivelicensetoSpringerNatureSwitzerland AG2023 Thisworkissubjecttocopyright.AllrightsaresolelyandexclusivelylicensedbythePublisher,whether thewholeorpartofthematerialisconcerned,specificallytherightsoftranslation,reprinting,reuse ofillustrations,recitation,broadcasting,reproductiononmicrofilmsorinanyotherphysicalway,and transmissionorinformationstorageandretrieval,electronicadaptation,computersoftware,orbysimilar ordissimilarmethodologynowknownorhereafterdeveloped. Theuseofgeneraldescriptivenames,registerednames,trademarks,servicemarks,etc.inthispublication doesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfromtherelevant protectivelawsandregulationsandthereforefreeforgeneraluse. Thepublisher,theauthors,andtheeditorsaresafetoassumethattheadviceandinformationinthisbook arebelievedtobetrueandaccurateatthedateofpublication.Neitherthepublishernortheauthorsor theeditorsgiveawarranty,expressedorimplied,withrespecttothematerialcontainedhereinorforany errorsoromissionsthatmayhavebeenmade.Thepublisherremainsneutralwithregardtojurisdictional claimsinpublishedmapsandinstitutionalaffiliations. ThisSpringerimprintispublishedbytheregisteredcompanySpringerNatureSwitzerlandAG Theregisteredcompanyaddressis:Gewerbestrasse11,6330Cham,Switzerland Preface Computer vision is a field of computer science that works on enabling computers to see, identify, and process images in the same way that human vision does, and then provide appropriate output. It is like imparting human intelligence and instincts to a computer. It is an interdisciplinary field that trains computers to interpretandunderstandthevisualworldfromdigitalimagesandvideos.Themain objectiveofthiseditedbookistoaddressanddisseminatestate-of-the-artresearch and development in the applications of intelligent techniques for computer vision. Thisbookprovidescontributionswhichincludetheory,casestudies,andintelligent techniquespertainingtothecomputervisionapplications.Thiswillhelpthereaders tograsp the extensive point of view and the essence of the recent advances inthis field. The prospective audience would be researchers, professionals, practitioners, andstudentsfromacademiaandindustrywhoworkinthisfield. Wehopethechapterspresentedwillinspirefutureresearchbothfromtheoretical and practical viewpoints to spur further advances in the field. A brief introduction toeachchapterisasfollows. Chapter1discussesthemachinelearningapproachesappliedtoautomaticsports video summarization. Chapter 2 proposes a new technique for lecture video seg- mentation and key frame extraction. The results are compared against six existing state-of-the-arttechniquesbasedoncomputationaltimeandshottransitions. Chapter 3 presents a system to detect the potholes in the pathways/roadways using machine learning and deep learning approaches. It uses HOG (histogram of oriented gradients) and LBP (local binary pattern) features to enhance the classificationalgorithmsperformance. Chapter 4 aims to explore various feature extraction techniques and shape detection approaches required for image retrieval. It also discusses the real-time applications of shape feature extraction and object recognition techniques with examples. Chapter5describesanapproachfortextureimageclassificationbasedonGray Level Co-occurrence Matrix (GLCM) features and machine learning algorithm. Chapter6presentsanoverviewofunimodalandmultimodalaffectivecomputing.It alsodiscussesthevariousmachinelearninganddeeplearningtechniquesforaffect recognition. Chapter 7 proposes a deep learning model for content-based image v vi Preface retrieval. It uses K-Means clustering algorithm and Hamming distance for faster retrievaloftheimage. Chapter 8 provides a bio-inspired convolutional neural network (CNN)-based model for COVID-19 diagnosis. A cuckoo search algorithm is used to improve theperformanceoftheCNNmodel.Chapter9presentsconvolutionalCapsNetfor detectingCOVID-19diseaseusingchestX-rayimages.Themodelobtainsfastand accuratediagnosticresultswithlesstrainableparameters. Chapter 10proposes adeep learning frameworkforanautomated hand gesture recognition system. The proposed framework classifies the input hand gestures, each represented by a set of histogram-oriented gradient feature vector into some predefinednumberofgestureclasses. Chapter11presentsanewhierarchicaldeeplearningbasedapproachforseman- tic segmentation of 3D point cloud. It involves nearest neighbor search for local featureextractionfollowedbyanauxiliarypretrainednetworkforclassification. Chapter 12 summarizes that the proposed model acts as a better automatic colorization for colored and grayscale images without human intervention. The proposed model predicted the color for the new images with good prediction accuracyclosetotherealimages.Infuture,suchautomaticcolorizationtechniques helptoidentifyvintageimagesormovieswithgrayscaleimageswiththeirdetails inaveryclearmanner. Chapter 13 proposes a generative adversarial network (GAN) for hyperspectral image classification. It uses dynamic mode decomposition (DMD) to reduce the redundantfeaturesinordertoattainbetterclassification.Chapter14presentsabrief introduction about the methodologies used for identifying diabetic retinopathy. It alsousesconvolutionalneuralnetworkmodelstoachieveaneffectiveclassification fordiabeticdetectionofretinalfundusimages. Chapter 15 proposes a modified differential evolution (DE), best neighborhood DE(BNDE), to solve discrete-valued benchmarking and real-world optimization problems.Theproposedalgorithmincreasestheexploitationandexplorationcapa- bilitiesoftheDEandtoreachtheoptimalsolutionfaster.Inaddition,theproposed algorithmisappliedtograyscaleimageenhancement. Chapter 16 presents an overview of the main swarm-based solutions proposed to solve problems related to computer vision. It presents a brief description of the principles behind swarm algorithms, as well as the basic operations of swarm methodsthathavebeenappliedincomputervision. Wearegratefultotheauthorsandreviewersfortheirexcellentcontributionsfor making this book possible. Our special thanks go to Mary James (EAI/Springer InnovationsinCommunicationandComputing)fortheopportunitytoorganizethis editedvolume. Preface vii We are grateful to Ms. Eliška Vlcˇková (Managing Editor at EAI – European AllianceforInnovation)fortheexcellentcollaboration. We hope the chapters presented will inspire researchers and practitioners from academiaandindustrytospurfurtheradvancesinthefield. Coimbatore,TamilNadu,India B.VinothKumar Coimbatore,TamilNadu,India P.Sivakumar Puducherry,Karaikal,India B.Surendiran Denton,TX,USA JunhuaDing January2023 Contents ASystematicReviewonMachineLearning-BasedSportsVideo SummarizationTechniques ..................................................... 1 VaniVasudevanandMohanS.Gounder ShotBoundaryDetectionfromLectureVideoSequencesUsing HistogramofOrientedGradientsandRadiometricCorrelation........... 35 T. Veerakumar, Badri Narayan Subudhi, K. Sandeep Kumar, NikhilO.F.DaRocha,andS.Esakkirajan DetectionofRoadPotholesUsingComputerVisionandMachine LearningApproachestoAssisttheVisuallyChallenged..................... 61 U.AkshayaDeviandN.Arulanand ShapeFeatureExtractionTechniquesforComputerVision Applications....................................................................... 81 E.Fantin IrudayaRajandM.Balaji GLCM Feature-Based Texture Image Classification Using MachineLearningAlgorithms ................................................. 103 R.Anand,T.Shanthi,R.S.Sabeenian,andS.Veni ProgressinMultimodalAffectiveComputing:FromMachine LearningtoDeepLearning ..................................................... 127 M.ChanchalandB.VinothKumar Content-Based Image Retrieval Using Deep Features andHammingDistance.......................................................... 151 R. T.AkashGunaandO. K.Sikha BioinspiredCNNApproachforDiagnosingCOVID-19Using ImagesofChestX-Ray .......................................................... 181 P. Manju Bala, S. Usharani, R. Rajmohan, T. Ananth Kumar, andA.Balachandar ix x Contents InitialStageIdentificationofCOVID-19UsingCapsuleNetworks ........ 203 ShamikaGanesan,R.Anand,V.Sowmya,andK.P.Soman DeepLearninginAutoencoderFrameworkandShapePriorfor HandGestureRecognition...................................................... 223 Badri Narayan Subudhi, T. Veerakumar, Sai Rakshit Harathas, RohanPrabhudesai,VenkatanareshbabuKuppili,andVinitJakhetiya Hierarchical-BasedSemanticSegmentationof3DPointCloud UsingDeepLearning ............................................................ 243 J.Narasimhamurthy,KarthikeyanVaiapury,RamanathanMuthuganapathy, andBalamuralidharPurushothaman ConvolutionNeuralNetworkandAuto-encoderHybridScheme forAutomaticColorizationofGrayscaleImages............................. 253 A.Anitha,P.Shivakumara,ShreyanshJain,andVidhiAgarwal DeepLearning-BasedOpenSetDomainHyperspectralImage ClassificationUsingDimension-ReducedSpectralFeatures ................ 273 C. S.Krishnendu,V.Sowmya,andK. P.Soman An Effective Diabetic Retinopathy Detection Using Hybrid ConvolutionalNeuralNetworkModels........................................ 295 NiteeshKumar,RashadAhmed,B.H.Venkatesh,M.AnandKumar ModifiedDiscreteDifferentialEvolutionwithNeighborhood ApproachforGrayscaleImageEnhancement................................ 307 AnishaRadhakrishnanandG.Jeyakumar Swarm-BasedMethodsAppliedtoComputerVision........................ 331 María-LuisaPérez-Delgado Index............................................................................... 357 A Systematic Review on Machine Learning-Based Sports Video Summarization Techniques VaniVasudevan andMohanS.Gounder 1 Introduction Sports video summarization is one of the interesting fields of research as it tends to generate a highlight of the broadcast video. Usually, the broadcasted sports videosarelonger,andtheaudiencesmaynothaveenoughtimetowatchtheentire durationofthegame.Someofthesportslikesoccer(football),basketball,baseball, tennis, golf, cricket, and rugby are played for the duration of 90–180 minutes per match.Hence,creatingasummarizationthatcontainsonlyeventsandexcitements ofinterestpertainingtoindividualsportsisanintensehumantask. There are several learning and non-learning-based techniques in the literature that attempt to automate the process of creating such highlights or summarization video. In addition to it, in recent years, the advancements of deep learning techniques have also contributed to accomplish remarkable results in the sports videosummarization.Figure1showsthegrowingnumberofpublicationsthatare associatedwith“sportsvideosummarization”overthepasttwodecades. Video summarization techniques [2] have been widely used in many types of sports. The choice of sports/games chosen for this systematic review are based on thefollowingcriteria:(1)sportswithhighaudiencebase(https://www.topendsports. com/world/lists/popular-sport/fans.html),(2)sportswherethesponsorshipismore, (3) sports with more watch views/hours, (4) sports where the research potential is highwithlargedatasets,(5)sportswheretheneedfortechnologicaladvancementis veryhigh,(6)frequencyoftheoccurrenceofthegame/sportinayear,(7)number of countries participating in the sports, and (8) number of countries hosting the V.Vasudevan DepartmentofCSE,NitteMeenakshiInstituteofTechnology,Bengaluru,India M.S.Gounder((cid:2)) DepartmentofISE,NitteMeenakshiInstituteofTechnology,Bengaluru,India ©TheAuthor(s),underexclusivelicensetoSpringerNatureSwitzerlandAG2023 1 B.V.Kumaretal.(eds.),SmartComputerVision,EAI/SpringerInnovationsin CommunicationandComputing,https://doi.org/10.1007/978-3-031-20541-5_1

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.