Hi!I'mYushiroYamashita,adatascientistatHACARUS.Today,Iwillliketointroducetheapplicationofsparsemodelingtopointclouddata. PointCloudIwanttofirststartbyintroducingwhatpointcloudis.Pointcloudisathree-dimensionaldatainwhichthecoordinatevaluesofeachpointarerecordedinaformatsuchas(x,y,z).Unlikeimageandvideodata,pointcloudhasauniquecharacteristicthatthedataarenotarrangedregularlyat…
Hello,I'mHaruyukiTago,EdgeEvangelistatHACARUSTokyoR&DCenter.Inthisseriesofarticles,IwillsharesomeinsightsfrommydecadeslongexperienceinthesemiconductorindustryandcommentondifferentAIindustryrelatedtopicsfrommyuniqueperspective.Sincethisisthefirsttime,I'llstartbyintroducingmyself.In1977,IjoinedthesemiconductordivisionofTokyoShibauraElectricCo.andwasinvolvedinthedesignanddevelopmentoftheEmotionEngine,theCPUin…
ThisisRyujiMasui,DataScientistatHACARUS.Inthisarticle,I'mgoingtofocusonhowtoapplySparseModelingtoHyperspectralImaging,atechnologyHACARUShasstronginterestin.HyperspectralImageLet'sstartoffbyexplainingwhatHyperspectralImages(HSI)are.NormalcamerasthatarefamiliartousarecalledRGBcameras,basedliterallyonthethreeprimarycolorsoflight:red,green,andblue.Suchcamerassenseeachwavelengthtocaptureimages.Insimplerterms,bymixing…
TL;DR-CVATwasusedasanannotationtoolformedicalAIdevelopment-CVATwasadjustedtobecompatibletoreadDICOM-AclasstoreadinDICOMwasaddedtocvat/apps/engine/media_extractors.py-Somefeaturesthatwillbeusefulinrealprojects,suchasdefaultlabelsettingswerealsoadded1.AboutCVATHello,mynameisYuyaSumie,adatascientistatHACARUS.Inordertodevelopmachinelearningalgorithms,weneedtoknowhowtofinalize…
Today'sblogwillbedeliveredbyYoshinobuOgura,DataScientistatHACARUS.Hello,mynameisYoshinobuOgura,andIamaHACARUSDataScientist.Inthisarticle,Iwillbeusingdeeplearningtoestimatehumanskeletalstructure-knownas"HumanPoseEstimation".AboutHumanPoseEstimationThefinalgoalaimofthisarticleistocreateabaselineforHumanPoseEstimation.HumanPoseEstimationisatechniquetomodelthehumanskeletonfromimages,whichis…
Inthisarticle,TakahiroSuzuki,adatascientistatHACARUSexplainstheconvexoptimizationproblemcoveringallthebasictermsyouneedtounderstandit.0.IntroductionMynameisTakahiroSuzuki,datascientistintheHACARUSTohokuteam.Asmentionedinapreviousblog, HACARUShasstartedstudysessionstolearnmoreaboutsparsemodelingaspartofoureverydaybusiness.TheSparseModelingStudySessionsarecontributingtothepromotionofmutualknowledgeexchangeamongHACARUSmembers-…
WhatisDictionaryLearning?Indictionarylearning,patchesofsignalssuchasimagesandwaveformsarerepresentedasaweightedlinearsumofdictionaries(clustersofatoms)(Figure1).Thisallowstheextractionofessentialsignalsandavarietyofapplicationsarepossibleaslistedbelow:・Noisesuppression(Figure1)・Super-Resolution・DataDeficiencyCompletion・DataCompression・Classification,etc.Inadditiontoawiderangeofapplications,oneappealingfeatureistheabilitytolearnandpredictusingimagesofdifferent…
Inthisarticle,DaisukeTakahashi,adatascientistatHACARUSexplainstheprocessofHumanPartSegmentationusingahighperformanceedgedevice;theJetsonNanoDeveloperKit. WhatistheJetsonNanoDeveloperKit?TheJetsonNanoisapurposebuiltedgedevicebyNVIDIAthatallowsforGPUbasedoperationsinalightandcosteffectivepackage.Muchlikeastandardcomputeritallowsyoutorunvariousapplications.ForthisusecaseweinstalledaLinux…
AspartofourongoingworktointegratethelatestindustrialrobotswithHACARUSAItechnologieswe'vejusttaughtourresidentMitsubishiElectricRobottorecognizehumans,andtorespondrespectfullytoahumanbow-withabow!Checkitoutbelow:https://vimeo.com/397903246
Clicktheimagebelowforfullstory-
