AI LabAI Lab

2020

7/

21

Hyperspectral Imaging and Sparse Modeling – Practical Experiments

ThisisRyujiMasui,DataScientistatHACARUS.Inthisarticle,I'mgoingtofocusonhowtoapplySparseModelingtoHyperspectralImaging,atechnologyHACARUShasstronginterestin.HyperspectralImageLet'sstartoffbyexplainingwhatHyperspectralImages(HSI)are.NormalcamerasthatarefamiliartousarecalledRGBcameras,basedliterallyonthethreeprimarycolorsoflight:red,green,andblue.Suchcamerassenseeachwavelengthtocaptureimages.Insimplerterms,bymixing…

2020

7/

07

Human Pose Estimation Experiment

Today'sblogwillbedeliveredbyYoshinobuOgura,DataScientistatHACARUS.Hello,mynameisYoshinobuOgura,andIamaHACARUSDataScientist.Inthisarticle,Iwillbeusingdeeplearningtoestimatehumanskeletalstructure-knownas"HumanPoseEstimation".AboutHumanPoseEstimationThefinalgoalaimofthisarticleistocreateabaselineforHumanPoseEstimation.HumanPoseEstimationisatechniquetomodelthehumanskeletonfromimages,whichis…

2020

6/

11

Sparse Modeling Study Sessions: A Convex Optimization Approach to LASSO – What is the Convex Optimization Problem?

Inthisarticle,TakahiroSuzuki,adatascientistatHACARUSexplainstheconvexoptimizationproblemcoveringallthebasictermsyouneedtounderstandit.0.IntroductionMynameisTakahiroSuzuki,datascientistintheHACARUSTohokuteam.Asmentionedinapreviousblog, HACARUShasstartedstudysessionstolearnmoreaboutsparsemodelingaspartofoureverydaybusiness.TheSparseModelingStudySessionsarecontributingtothepromotionofmutualknowledgeexchangeamongHACARUSmembers-…

2020

5/

22

A peek inside our Sparse Modeling Study Sessions: Image Analysis

WhatisDictionaryLearning?Indictionarylearning,patchesofsignalssuchasimagesandwaveformsarerepresentedasaweightedlinearsumofdictionaries(clustersofatoms)(Figure1).Thisallowstheextractionofessentialsignalsandavarietyofapplicationsarepossibleaslistedbelow:・Noisesuppression(Figure1)・Super-Resolution・DataDeficiencyCompletion・DataCompression・Classification,etc.Inadditiontoawiderangeofapplications,oneappealingfeatureistheabilitytolearnandpredictusingimagesofdifferent…

2020

4/

23

Evaluation of Human Part Segmentation using the Jetson Nano Developer Kit

Inthisarticle,DaisukeTakahashi,adatascientistatHACARUSexplainstheprocessofHumanPartSegmentationusingahighperformanceedgedevice;theJetsonNanoDeveloperKit. WhatistheJetsonNanoDeveloperKit?TheJetsonNanoisapurposebuiltedgedevicebyNVIDIAthatallowsforGPUbasedoperationsinalightandcosteffectivepackage.Muchlikeastandardcomputeritallowsyoutorunvariousapplications.ForthisusecaseweinstalledaLinux…

2020

3/

16

Greeting a New Member @KYOTO HQ

AspartofourongoingworktointegratethelatestindustrialrobotswithHACARUSAItechnologieswe'vejusttaughtourresidentMitsubishiElectricRobottorecognizehumans,andtorespondrespectfullytoahumanbow-withabow!Checkitoutbelow:https://vimeo.com/397903246

2020

2/

12

HACARUS featured in Ledge.ai about Explainable AI

Clicktheimagebelowforfullstory-

2019

12/

26

HACARUS Gives Lecture at Shiga University Graduate School of Data Science

OnDecember19,2019,TakashiSomeda,HACARUS’ChiefTechnicalOfficer,NaokiKitora,itsChiefDataScientistandRyujiMasui,DataScientistgavealectureatShigaUniversity,totheGraduateSchoolofDataScience,introducingrealworldexamplesofutilizingmachinelearningforimprovingvisualinspectionatmanufacturingsites.TheGraduateSchoolofDataScienceatShigaUniversity,whichopenedinAprilofthisyear,hasamixedstudentbody,consistingofprofessionals,inadditiontothosewhohaveadvanced…

2019

10/

25

Hacarus @ PyCon Taiwan 2019

ThisyearHacarussponsoredPYCONTaiwanaspartofourongoingengagementwithwithPYCONcommunity,belowfollowsaconference reportwrittenbyNinzEclarin,oneofourEngineersattheManilaDevelopmentCenter. I'vebeentosomePyConeventsinAsiabefore,butPyConTaiwanstrikessomeawesomedistinction.Imean,whocansaynotoagreatselectionoffoodanddrinksprovidedatthevenue?UnlimitedMilkTeaforthewin!Butasidefromtheawesomefood,PyCon…

2019

9/

10

Part 3 of Sparse Modeling for IT Engineers on CodeZine

ThethirdinstallmentofourseriesonSparseModelingforITengineerscameouttodayinCodeZine.Thistime,itiswrittenbyIppeiUsamiofHacarusDataScienceTeam-Readthefullstoryinthelinkbelow(onlyJapanese) スパースモデリングのモデルを評価する～LASSO推定値の評価方法