AA?Efficientfruitselectionbynon-destructiveevaluationoffruitqualityinpost-harvestmanagement?Phenotypingofdataanddeeplearning?Quantitativeanalysisoforchardsusingimagedataforhighqualityfruitproduction20YutaroOSAKOAssistantProfessor, Ph. DResearch keywords:?Fruitand fruit tree?Nondestructiveevaluation?Image phenotyping?Deep learning“Pomology”isabranchofhorticulturalsciencethatfocusesonfruitandfruittrees.Thisfieldinvolvesbasicresearchintophysiologyoffruitandfruittrees,aswellasappliedresearchnecessaryfortheefficientfruitproduction.Inparticular,ourresearchaimstodevelopnewapproachestoefficientfruitproduction.Wearegoingtodevelopnon-destructivephenotypingtechniquesusingdeeplearning(AI:ArtificialIntelligence),basedonfruitandorchardimages.Wealsoaimtoquantifyandanalyzecultivationtechniquesthatenablethehighqualityfruitproductionimagedataanalysis.fruittraitsusingimageAnexampleofamodelfordiscriminatingbetweensmall-seededfruitandlarge-seededfruit.Althoughthereisnodifferenceinappearancebetweenthetwofruits,deeplearningmodelsuggeststhatthefeaturesmaybepresentinthefruitapexandpeduncleside.Aexampleofadeeplearningmodelthatcaneasilydetectpersimmonfruitinimagesandvideos.Itcanleadtoyieldpredictionandmoreefficientcultivationmanagement.Bioresource ScienceDivisionLaboratory of Pomology
元のページ ../index.html#24