I-Google iphakamise ukuthi sikwazi thatha izithombe ezinhle kakhulu bese uqopha amavidiyo amahle kakhulu Futhi ngalokhu, eminyakeni yamuva nje yenze ukuthi kutholakale kubasebenzisi imisebenzi eminingi njengethuluzi lokuzenzakalela elimhlophe kuzithombe ze-Google, ithuthukise izixazululo zokuqiniswa kwesithombe samavidiyo futhi yethule ngisho nethuluzi elisivumela ukuthi sihlanganise izithombe zasendulo.
Kepha manje inkampani yenjini yokusesha ifuna ukuqhubekela phambili futhi isebenzisa ubuhlakani bokufakelwa ezithombeni ezakha ifomu le- thinta kabusha futhi uthuthukise izithombe zakho ngaphambi kokuthi uzithathe.
Ososayensi abavela eMIT nakuGoogle basebenzisana ku- sebenzisa ama-algorithm wokufunda ngomshini ukuze uthuthukise izithombe ngesikhathi sangempela ukuwabonisa esikrinini se-smartphone. Kodwa kukhona nokunye okuningi ngoba akukhulunywa ngokulungiswa okuzenzakalelayo okusetshenziswe ngendlela efanayo kuzo zonke izithombe, njengoba kwenza ezinye izinhlelo eziningi, kodwa kunalokho izithuthukisi zenzelwe izithombe ngazinye, ngokuya ngezimo zabo ezithile.
Ukufeza lokhu, iqembu "seliqeqeshe" lawa manethiwekhi we-neural anezithombe ezingaphezu kwe-5.000 ezazitholwe kabusha ngabathwebuli bezithombe abahlanu abahlukene. Ngenxa yalokhu I-Artificial Intelligence ikwazile ukukhiqiza ifomula okukuvumela ukuthi usebenze ezithombeni ezithinta kabusha ngokusebenzisa izilungiselelo ezifanele kakhulu zesithombe ngasinye.
Ungabona isibonelo esithombeni esifakiwe salokhu okuthunyelwe, nakwividiyo elandelayo ethunyelwe nguMichael Gharbi:
Le software ingasebenza kuma-Smartphones ane- isikhathi esincane sokubambezeleka futhi futhi nge ukusetshenziswa kwebhethri okuncane, izinto ezimbili kuze kube manje sezibambezele ukusetshenziswa kwalokhu kucubungula isithombe ezingcingweni.
Okwamanje, ayikho iGoogle noma i- MIT Bathuthukise isikhathi esingahle senziwe sokuthengiswa kwalobu buchwepheshe kepha, kubonakala sengathi ingafinyelela i-Android esikhathini esizayo esingekude kakhulu.