Kamfanin Beeline ya kirkiro wata hanyar sadarwa ta musamman wacce za ta taimaka wajen neman mutanen da suka bata: ana kiran dandalin "Beeline AI - Neman Mutane."
An tsara maganin don sauƙaƙe aikin ƙungiyar bincike da ceto." Tun daga shekarar 2018, wannan tawagar tana amfani da jirage marasa matuka don gudanar da bincike a cikin dazuzzuka da wuraren masana'antu na birane. Koyaya, nazarin hotunan da aka samu daga kyamarori marasa matuƙa yana buƙatar shigar da yawan masu sa kai. Bugu da ƙari, wannan yana ɗaukar lokaci mai yawa.
Cibiyar sadarwa ta jijiyoyi "Beeline AI - Binciken Mutane" an tsara shi don sarrafa tsarin sarrafa hoto. Ana da'awar cewa ƙwararrun algorithms na iya rage lokacin dubawa da rarraba hotuna da aka karɓa sau biyu da rabi.
Dandalin yana amfani da fasahohin hanyar sadarwa na jujjuyawar jijiyoyi, wanda ke ƙara ingantaccen kayan aikin hangen nesa na kwamfuta. An horar da cibiyar sadarwar jijiyoyi akan ainihin tarin hotuna. Gwaje-gwaje sun nuna cewa daidaiton samfurin akan hotunan gwajin yana kusa da 98%.
Babban aikin "Beeline AI - Bincike na Mutane" shine warware "marasa amfani" da hotuna marasa fahimta waɗanda ba su da mutum ko halayen da ke nuna cewa akwai mutum a wannan wuri. Wannan yana ba ƙungiyar masu bincike damar mayar da hankali nan da nan kan abubuwan da za su iya tasiri.
Tsarin zai iya dacewa da yanayi daban-daban. Hakanan yana samun daidaitattun abubuwa duka daga tsayin mita 30-40 da kuma tsayin jirgin sama na mita 100. A lokaci guda, cibiyar sadarwar jijiyoyi tana da ikon sarrafa hotuna tare da babban matakin gani "amo" - bishiyoyi, shimfidar yanayi, faɗuwar rana, da dai sauransu.
"Mai yiwuwa, cibiyar sadarwar jijiyoyi tana da ikon gano mutane da abubuwa a duk wuraren bincike, kamar gandun daji, swamps, filayen, birane, ba tare da la'akari da lokacin shekara da tufafin mutum ba, tun da algorithm an saita shi don yin aiki a kowane lokaci. shekarar kuma za a iya gane matsayin da ba daidai ba a sararin samaniya, alal misali, mutumin da ke zaune, kwance ko wani sashi ya rufe shi," in ji Beeline.
source: 3dnews.ru