Paper Title
Skin Cancer Detection using Machine Learning

Abstract
SkinTisTanTextraTordinaryThumanTstructure.TThereTareTvariousTtypesTofTskinTcancer.AndTpeopleTareTunawareTaboutTit.TThereTareTlargeTnumberTofTskinTdiseasesTandTsomeTofthemTareTmostTcommon.TDueTtoTlackTofTmedicalTfacilitiesTavailableTInTremoteTareas,patientsTusuallyTignoreTit.TTheTdiagnoseTofthisTskinTcanceralsoTtakeTtheTlongertime.MelanomaTskinTcancerTdetectionTatTanTearlyTstageTisTcrucialTforTanTefficientTtreatment.TRecently,TitTisTwellknownTthat,TtheTmostTdangerousTformTofTskinTcancerTamongTtheTotherTtypesTofTskinTcancerTisTmelanomaTbecauseTit’sTmuchTmoreTlikelyTtoTspreadTtoTotherTpartsTofTtheTbodyTifnotTdiagnosedTandTtreatedTearly.TTheTnonTinvasiveTmedicalTcomputerTvisionTorTmedicalTimageTprocessin plays increasingly significantTroleTinTclinical diagnosisTofTdifferentTdiseases.TsuchTtechniquesTprovideTanTautomaticTimageTanalysisTtoolTforTanTaccurateTand fastTevaluationTofTtheTlesion.TTheTstepsTinvolvedTinTthisTstudyTareTcollectingTDermoscopyTimageTdatabase,Tpreprocessing,TsegmentationTusingTthresholding,TstatisticalTfeatureTextractionTusingTGrayTLevelTCooccurrenceTMatrixT(GLCM),TAsymmetry,TBorder,TColor,TDiameter,T(ABCD)Tetc.,TfeatureTselectionTusingTPrincipalTcomponentTanalysisT(PCA),TcalculatingTtotalTDermoscopyTScoreTandTthenTclassificationTusingTConvocationTneuralTnetworkT(CNN).TresultsTshowTthatTtheTachievedTclassificationTaccuracyTisT92.1.