Plasmodium gallinaceum is an important causative agent of avian malaria, which causes more than 80% mortality if left untreated. The disease ...SkiptomaincontentThankyouforvisitingnature.com.YouareusingabrowserversionwithlimitedsupportforCSS.Toobtainthebestexperience,werecommendyouuseamoreuptodatebrowser(orturnoffcompatibilitymodeinInternetExplorer).Inthemeantime,toensurecontinuedsupport,wearedisplayingthesitewithoutstylesandJavaScript.AdvertisementnaturescientificreportsarticlesarticleClassificationforavianmalariaparasitePlasmodiumgallinaceumbloodstagesbyusingdeepconvolutionalneuralnetworksDownloadPDFSubjectsMathematicsandcomputingParasitologyPathologyAbstractTheinfectionofanavianmalariaparasite(Plasmodiumgallinaceum)indomesticchickenspresentsamajorthreattothepoultryindustrybecauseitcauseseconomiclossinboththequalityandquantityofmeatandeggproduction.Computer-aideddiagnosishasbeendevelopedtoautomaticallyidentifyavianmalariainfectionsandclassifythebloodinfectionstagedevelopment.Inthisstudy,fourtypesofdeepconvolutionalneuralnetworks,namelyDarknet,Darknet19,Darknet19-448andDensenet201areusedtoclassifyP.gallinaceumbloodstages.Werandomlycollectedadatasetof12,761single-cellimagesconsistingofthreeparasitestagesfromten-infectedbloodfilmsstainedbyGiemsa.Allimageswereconfirmedbythreewell-trainedexaminers.Thestudymainlycomparedseveralimageclassificationmodelsandusedbothqualitativeandquantitativedatafortheevaluationoftheproposedmodels.Inthemodel-wisecomparison,thefourneuralnetworkmodelsgaveushighvalueswithameanaverageaccuracyofatleast97%.TheDarknetcanreproduceasuperiorperformanceintheclassificationoftheP.gallinaceumdevelopmentstagesacrossanyothermodelarchitectures.Furthermore,theDarknethasthebestperformanceinmultipleclass-wiseclassification,withaveragevaluesofgreaterthan99%inaccuracy,specificity,andsensitivity.Italsohasalowmisclassificationrate(< 1%)thantheotherthreemodels.Therefore,themodelismoresuitableintheclassificationofP.gallinaceumbloodstages.Thefindingscouldhelpuscreate