Research Article | Open Access
DIABETIC RETINOPATHY ANALYSIS
Mrs B MAHALAKSHMI, DHANWADA ANISHA, NAKKA SHREYA, DEEKONDA ARAVIND, KATIKAM UMESH REDDY
Pages: 750-756
Abstract
Agreatchallengeinthebiomedicalengineeringisthenon-invasiveassessmentofthephysiologicalchangesoccurringinsidethehumanbody.Specifically,detectingthe abnormalitiesinthehumaneyeisextremelydifficultduetothevariouscomplexitiesassociatedwiththeprocess.Retinalimagescapturedbydigitalcamerascanbeusedtoidentifythenatureoftheabnormalitiesaffectingthehumaneye.Conventionaldiseaseidentificationtechniquesfromretinalimagesaremostlydependentonmanualintervention.Humanobservationishighlypronetoerror,thesuccessrateofthesetechniquesisquitelow.DiabeticRetinopathyisonesuchdiseaseofretinawhichoccursinpeoplesufferingfromlongstandingdiabetes.ItisamultistageprogressingdiseasenamelyNDPR andPDR. Microaneurysms,hemorrhages andexudatesaretheabnormalfeaturescommonlyobservedintheretinalimageofapersonaffectedbydiabeticretinopathy.Imageprocessingtechniquesareusedtopre-processthefundusimage,whichisfollowedbysegmentationofanomalies.Featureextractionisdoneandthedetectedfeaturesareusedtoclassifythedifferentstagesofdiabeticretinopathy.TheclassificationtechniqueusedisFast Regional convolutional Neural Network,andtheaccuracyobtainedis98.9%.
Keywords
Agreatchallengeinthebiomedicalengine
