I-biometrics izilinganiso zomzimba nezibalo ezihlobene nezici zomuntu. Ukuqinisekiswa kwe-biometric (noma ukuqinisekiswa okungokoqobo) kusetshenziswa kwisayensi yekhompiyutha njengendlela yokuhlonza nokulawula ukufinyelela. Iphinde isetshenziselwe ukukhomba abantu ngabanye emaqenjini agadiwe.

Izihlonzi ze-biometric yizici ezihlukile, ezilinganisekayo ezisetshenziselwa ukulebula nokuchaza abantu ngabanye. Izihlonzi zebhayomethrikhi zivame ukuhlukaniswa njengezici zefiziyoloji ezihlobene nokuma komzimba. Izibonelo zifaka, kodwa azikhawulelwe kuzigxivizo zeminwe, imithambo yesundu, ukubona ubuso, i-DNA, ukuphrinta kwesundu, ijometri yesandla, ukubonwa kwe-iris, i-retina, nephunga/iphunga.

Ubuchwepheshe bokuhlonza i-biometric buhlanganisa isayensi yekhompiyutha, i-optics kanye ne-acoustics nezinye isayensi yemvelo, isayensi yebhayoloji, izimiso ze-biosensors kanye ne-biostatistics, ubuchwepheshe bezokuphepha, nobuchwepheshe bobuhlakani bokwenziwa nezinye isayensi eminingi eyisisekelo kanye nobuchwepheshe bokusebenzisa obusha. Kuyizixazululo zobuchwepheshe ezihlanganisa imikhakha eminingi.

Eminyakeni yakamuva, ngokuthuthukiswa kobuhlakani bokwenziwa, ubuchwepheshe bokuhlonza i-biometric buye bavuthwa kakhulu. Njengamanje, ubuchwepheshe bokubona ubuso bumelela kakhulu i-biometrics.

Ukubonwa kobuso

Inqubo yokuqashelwa kobuso ihlanganisa ukuqoqwa kobuso, ukutholwa kobuso, ukukhipha isici sobuso nokubonwa kokufanisa ubuso. Inqubo yokuqaphela ubuso isebenzisa ubuchwepheshe obuhlukahlukene obufana ne-algorithm ye-AdaBoos, inethiwekhi ye-convolutional neural kanye nomshini wokusekelayo we-vector ekufundeni komshini.

Inqubo yokuqashelwa kobuso

Njengamanje, ubunzima bokubona ubuso bendabuko obuhlanganisa ukuzungezisa ubuso, ukuvala, ukufana, njll. buthuthukiswe kakhulu, okuthuthukisa kakhulu ukunemba kokubonwa kobuso. Ubuso obuyi-2D, ubuso be-3D, ubuso obunezibukeli eziningi Imodi ngayinye inezimo ezihlukile zokutholwa kokutholwa, idigri yokuphepha kwedatha nokuzwela kobumfihlo, njll., futhi ukungezwa kokufunda okujulile kwedatha enkulu kwenza i-algorithm yokuqaphela ubuso be-3D yengeze iziphambeko zokuqagela kwe-2D, Ingakwazi ukuhlonza ngokushesha ubunikazi bomuntu, okulethe impumelelo ethile ekusetshenzisweni kokuqashelwa kobuso obunezinhlangothi ezimbili.

Ngesikhathi esifanayo, ubuchwepheshe bokubona i-biometric okwamanje busetshenziswa njengobuchwepheshe obuyinhloko bokuthuthukisa ukuphepha kokuqashelwa kobuso, obungamelana ngokuphumelelayo nokukhwabanisa komgunyathi njengezithombe, amavidiyo, amamodeli e-3D, namamaski okwenziwa, futhi anqume ngokuzimela ubunikazi bomgunyathi. abasebenzisi abasebenzayo. Njengamanje, ngokuthuthuka okusheshayo kobuchwepheshe bokubona ubuso, izinhlelo zokusebenza eziningi ezintsha ezifana namadivayisi ahlakaniphile, ezezimali nge-inthanethi, nokukhokha ngobuso sekuye kwaduma kakhulu, okuletha isivinini kanye nokunethezeka empilweni nasemsebenzini wawo wonke umuntu.

Ukuqashelwa kwe-Palmprint

Ukuqashelwa kwe-Palmprint kuwuhlobo olusha lobuchwepheshe bokuqashelwa kwe-biometric, obusebenzisa i-palmprint yomzimba womuntu njengesici okuqondiswe kuso, futhi luqoqa ulwazi lwebhayoloji ngobuchwepheshe be-imaging be-multispectral. Ukuqashelwa kwe-multi-spectral palmprint kungathathwa njengemodeli yobuchwepheshe bokuqashelwa kwe-biometric obuhlanganisa izindlela eziningi nezici eziningi eziqondiwe. Lobu buchwepheshe obusha buhlanganisa izici ezintathu ezibonakalayo zesikhumba, ukuphrinta kwesundu kanye nemithambo yemithambo ukuze kunikeze ulwazi oluningi ngesikhathi esisodwa futhi kwandise ukuhlukaniswa kwezici eziqondiwe.

Kulo nyaka, ubuchwepheshe bokuqashelwa kwesundu e-Amazon, obuqanjwe ngekhodi u-Orville, sebuqalile ukuhlolwa. Isithwebuli siqala ngokuthola isethi yezithombe zangempela ze-infrared polarized, ezigxile ezicini zangaphandle zesundu, njengemigqa nokugoqa; lapho ithola isethi yesibili yezithombe ze-polarized futhi, igxila esakhiweni sesundu nezici zangaphakathi, njengemithambo, amathambo, izicubu ezithambile, njll. Izithombe eziluhlaza ziqale zicutshungulwe ukuze zinikeze isethi yezithombe eziqukethe izandla. Lezi zithombe zikhanyiswe kahle, zigxilile, futhi zibonisa intende ngomumo othile, endaweni ethile, futhi zilebulwe njengesandla sobunxele noma kwesokudla.

Njengamanje, ubuchwepheshe be-Amazon bokubona i-palmprint bungaqinisekisa ubuwena kanye nenkokhelo ephelele ngama-millisecond angu-300 kuphela, futhi abudingi abasebenzisi ukuthi babeke izandla zabo kudivayisi yokuskena, bavele bazulise futhi baskene ngaphandle kokuxhumana. Izinga lokuhluleka kwalobu buchwepheshe cishe li-0.0001%. Ngesikhathi esifanayo, ukuqashelwa kwe-palmprint ukuqinisekiswa kabili esigabeni sokuqala - okokuqala ukuthola izici zangaphandle, kanti okwesibili ukuthola izici zenhlangano zangaphakathi. Uma kuqhathaniswa nobunye ubuchwepheshe be-biometric mayelana nokuphepha, okuthuthukisiwe.

Ngaphezu kwezici ezingenhla zebhayomethrikhi, ubuchwepheshe bokuqashelwa kwe-iris bubuye budlondlobale. Izinga lokuqashelwa okungamanga lokubonwa kwe-iris liphansi njengo-1/1000000. Isebenzisa ikakhulukazi izici zokungaguquki kwempilo ye-iris kanye nomehluko ukukhomba ubunikazi.

Njengamanje, ukuvumelana embonini ukuthi ukuqashelwa kwendlela eyodwa kunezingqinamba kukho kokubili ukusebenza kokuqashelwa nokuphepha, futhi ukuhlanganiswa kwe-multi-modal kuyimpumelelo ebalulekile ekuqashelweni kobuso ngisho nokuqashelwa kwe-biometric-hhayi kuphela ngezinto eziningi Indlela ukuthuthukisa ukunemba kokubonwa kungaphinda kuthuthukise ukuguquguquka kwesigcawu kanye nokuvikeleka kobumfihlo bobuchwepheshe be-biometric ngokwezinga elithile. Uma kuqhathaniswa ne-algorithm yendabuko yemodi eyodwa, ingahlangabezana kangcono nezinga lezinga lezezimali lokuqashelwa okungamanga (eliphansi njengesigidi esisodwa esigidini esiyishumi), okuphinde kube inkambiso eyinhloko yokuthuthukiswa kokuhlonza i-biometric.

Isistimu ye-biometric ye-Multimodal

Amasistimu e-biometric amaningi asebenzisa izinzwa eziningi noma ama-biometrics ukuze anqobe imikhawulo yesistimu yebhayomethrikhi engashintshiwe.Isibonelo amasistimu wokuqaphela iris angafakwa engozini ngokuguga kwe-irises futhi ukuqashelwa kwezigxivizo zeminwe kungenziwa kubi kakhulu ngezigxivizo zeminwe ezigugile noma ezisikiwe. Nakuba amasistimu e-biometric e-unimodal enqunyelwe ubuqotho besihlonzi sawo, mancane amathuba okuthi amasistimu amaningana angaguquki abhekane nemikhawulo efanayo. Amasistimu e-biometric amaningi angathola amasethi olwazi kumaka ofanayo (okungukuthi, izithombe eziningi ze-iris, noma izikena zomunwe ofanayo) noma ulwazi oluvela ku-biometrics ehlukene (edinga ukuskena kwezigxivizo zeminwe, kusetshenziswa ukubonwa kwezwi, iphasikhodi ekhulunywayo).

Amasistimu e-biometric amaningi angahlanganisa lezi zinhlelo ze-unimodal ngokulandelana, ngesikhathi esisodwa, inhlanganisela yakho, noma uchungechunge, olubhekisela ezinhlotsheni ezilandelanayo, ezihambisanayo, ezilandelanayo nezilandelanayo, ngokulandelanayo.

I-CHANCCTVusungule uchungechunge lweama-lens e-biometricukubonwa kobuso, ukubonwa kwezigxivizo zesundu kanye nokuhlonza izigxivizo zeminwe kanye nokuhlonza iris.Isibonelo i-CH3659A iyilensi yokuhlanekezela engu-4k ephansi eyayidizayinelwe izinzwa ezingu-1/1.8''. Ifaka yonke imiklamo yengilazi necwecwe ene-11.95mm TTL nje. Ithwebula inkambu yokubuka evundlile engu-44 degrees. Le lensi ilungele ukubonwa kwe-palmprint.