Kavya Kopparapu, saurayi wanda ke zaune a Virginia, ya yi amfani da wayoyin komai da ruwanka, 3D firinta, da kuma fasaha ta wucin gadi zuwa ci gaba da ganewar ido. a šaukuwa tsarin bincike kuma mai arha don tantancewa maganin ciwon sukari. Wannan cuta cuta ce ta ciwon suga yana lalata jijiyoyin jini akan kwayar idon mutum kuma yana iya haifar da rashin gani.
Kopparapu, ta kasance mai sha'awar ilimin kimiyya a duk rayuwarta kuma bayan halartar taron shirye-shiryen shirye-shirye da Cibiyar Nazarin Mata da Fasahar Sadarwa ta organizedasa ta shirya, ta ƙara shirye-shirye a cikin abubuwan da take so.
Kakan Kopparapu, wanda ke zaune a Indiya, fara nuna alamun bayyanar na cutar a 2013. Zai iya faruwa sau da yawa ba tare da an sani ba, kuma kodayake daga ƙarshe an gano shi kuma an kula da shi, ganinsa ya tabarbare. A cewar Kopparapu, daga cikin jimillar Mutane miliyan 415 da ke fama da ciwon sukari, kashi daya bisa uku zai haifar da ciwon suga, kuma kodayake magani da tiyata na iya dakatarwa ko ma juya baya ga lalacewar ido idan an kama shi a kan lokaci, 50% ba za a bincikar lafiya ba, rabin marasa lafiya da siffofin masu tsanani za su makance cikin shekaru biyar.
"Rashin ganewar asali shine babban kalubale. A Indiya, akwai shirye-shirye da ke aika likitoci zuwa ƙauyuka da wuraren marasa galihu, amma akwai marasa lafiya da yawa kuma likitocin ido ne kawai.
Yayi tunanin ko akwai hanya mai sauki da mara tsada don gano cutar, kuma tunanin Eyagnosis ya fito, tsarin da zai iya juya doguwar tsada da tsada hanyar bincike mai sauki. Kopparapu ya fara aiki, ya dau lokaci mai yawa akan Google sannan ya aikawa likitoci da masu bincike sakonnin email, kafin ya tsara wani tsari. Ta haɗu tare da ɗan'uwanta kuma abokin karatunta, kuma yi amfani da hanyar sadarwa mai rikitarwa (CNN) don tabbatar da bincikar cutar AI a bayan Ciwon Ido. Cibiyoyin sadarwar yanar gizo bincika manyan bayanan bayanai kuma nemi irin waɗannan alamu, Tunda zane yayi kama da tsarin gani na kwakwalwar mutum, CNNs suna da kyau don rabewa.
Ya yi amfani da ResNet-50, wani CNN wanda masu bincike na Microsoft suka kirkira, don gina cibiyar sadarwar sa, kuma tayi amfani da 34.000 masu daukar ido samu a cikin database EyeGene daga Cibiyar Kiwon Lafiya ta Kasa a Amurka (NIH) kamar yadda bayanan koyo, don haka ita da ƙungiyarta za su iya koyar da tsarin AI don gane alamun cuta a cikin hotunan idanu kuma su ba da ganewar asali. Yawancin hotuna a cikin rumbun adana bayanan sun kasance marasa kyau ko kuma sun bayyana, amma a cewar Kopparapu, wannan daki-daki ya taimaka wajen inganta tsarin.
«Ingancin hotunan da aka yi amfani da su wajen koyon hanyar sadarwar jijiya yana wakiltar yanayin da za a samu ta hanyar amfani da wayar salula«
Tawagarsa sun horar da ResNet-50 zuwa gano cututtukan cututtukan cututtukan sukari kamar yadda ya dace a matsayin ainihin masanin ilimin cututtuka. Hakanan yana gano microaneurysms da jijiyoyin jini a cikin kowane hoto ba tare da buƙatar allurar fenti mai haske a cikin idon da aka gano ba.
Fallarshen ƙarshe, da Aditya Jyot Eye Asibiti a Mumbai ya yarda ya gwada aikin maganin cutar ido, kuma a cikin Nuwamba, ta aika samfurin 3D na farko da aka buga zuwa asibiti, kuma tsarin ya riga ya hyi cikakken bincike ga marasa lafiya biyar.
Ciwon ido yana da jan aiki a gaba wanda za'a iya binciko adadi mai yawa na cutar don tabbatar da cewa tsarin abin dogaro ne. Tsarin tabbatarwa wanda kowane aikin da ya danganci magani dole ne ya kasance yana da tsauri kuma tabbas zaiyi muku wahala samun babban kamfani don son taimaka muku. Amma duk waɗannan matsalolin ba sa hana ɗaukakar nasarar wannan matashi.