{"id":14756,"date":"2024-07-06T01:38:15","date_gmt":"2024-07-06T01:38:15","guid":{"rendered":"https:\/\/www.lifescienceart.com\/?p=14756"},"modified":"2024-07-06T01:38:15","modified_gmt":"2024-07-06T01:38:15","slug":"google-ai-image-localization-planet","status":"publish","type":"post","link":"https:\/\/www.lifescienceart.com\/hu\/science\/artificial-intelligence\/google-ai-image-localization-planet\/","title":{"rendered":"PlaNet: a Google pontos k\u00e9pfelismer\u0151 neur\u00e1lis h\u00e1l\u00f3zata"},"content":{"rendered":"<h2 class=\"wp-block-heading\">A Google \u00faj mesters\u00e9ges intelligenci\u00e1ja pontosan meghat\u00e1rozza a fot\u00f3k hely\u00e9t<\/h2>\n\n<h2 class=\"wp-block-heading\">\u00cdme a PlaNet: a Google k\u00e9pfelismer\u0151 neur\u00e1lis h\u00e1l\u00f3zata<\/h2>\n\n<p>A mesters\u00e9ges intelligencia (MI) ter\u00e9n a Google jelent\u0151s el\u0151rel\u00e9p\u00e9st tett a PlaNet fejleszt\u00e9s\u00e9vel, egy olyan neur\u00e1lis h\u00e1l\u00f3zattal, amely figyelemre m\u00e9lt\u00f3 pontoss\u00e1ggal k\u00e9pes meghat\u00e1rozni egy f\u00e9nyk\u00e9p hely\u00e9t. Ez az \u00e1tt\u00f6r\u00e9s forradalmas\u00edthatja a k\u00e9palap\u00fa alkalmaz\u00e1sokat, \u00e9s jav\u00edthatja a minket k\u00f6r\u00fclvev\u0151 vil\u00e1g meg\u00e9rt\u00e9s\u00e9t.<\/p>\n\n<h2 class=\"wp-block-heading\">A PlaNet m\u0171k\u00f6d\u00e9se<\/h2>\n\n<p>A PlaNet egy k\u00e9p pixeleit elemezve hat\u00e1rozza meg annak hely\u00e9t. A neur\u00e1lis h\u00e1l\u00f3zat betan\u00edt\u00e1s\u00e1hoz a kutat\u00f3k t\u00f6bb ezer f\u00f6ldrajzi \u201ecell\u00e1ra\u201d osztott\u00e1k a F\u00f6ldet, \u00e9s t\u00f6bb mint 100 milli\u00f3 geotaggelt k\u00e9pet vittek be. N\u00e9h\u00e1ny k\u00e9pet arra haszn\u00e1ltak, hogy megtan\u00edts\u00e1k a PlaNetnek, hogy egy adott cella melyik k\u00e9phez tartozik, m\u00edg m\u00e1sok az els\u0151dleges eredm\u00e9nyek ellen\u0151rz\u00e9s\u00e9re szolg\u00e1ltak.<\/p>\n\n<h2 class=\"wp-block-heading\">Leny\u0171g\u00f6z\u0151 pontoss\u00e1g<\/h2>\n\n<p>A tesztek sor\u00e1n a PlaNet leny\u0171g\u00f6z\u0151 eredm\u00e9nyeket \u00e9rt el. A k\u00e9pek 3,6 sz\u00e1zal\u00e9k\u00e1nak hely\u00e9t \u201eutcaszint\u0171 pontoss\u00e1ggal\u201d azonos\u00edtotta, 10,1 sz\u00e1zal\u00e9k\u00e1t v\u00e1rosi szinten, 28,4 sz\u00e1zal\u00e9k\u00e1t orsz\u00e1gos szinten, 48 sz\u00e1zal\u00e9k\u00e1t pedig kontinent\u00e1lis szinten. Ezek az eredm\u00e9nyek fel\u00fclm\u00falj\u00e1k az emberi teljes\u00edtm\u00e9nyt, hiszen a PlaNet helytelen tal\u00e1lgat\u00e1sai \u00e1tlagosan mind\u00f6ssze 702 m\u00e9rf\u00f6ldre voltak a t\u00e9nyleges helyt\u0151l, szemben az emberi alanyok 1400 m\u00e9rf\u00f6ldn\u00e9l is nagyobb t\u00e1vols\u00e1g\u00e1val.<\/p>\n\n<h2 class=\"wp-block-heading\">Alkalmaz\u00e1sok \u00e9s lehet\u0151s\u00e9gek<\/h2>\n\n<p>A PlaNet k\u00e9pess\u00e9geinek messze hat\u00f3 k\u00f6vetkezm\u00e9nyei vannak. Be\u00e9p\u00edthet\u0151 olyan eszk\u00f6z\u00f6kbe, mint a mobiltelefonok, hogy \u00f6sszetett k\u00e9pelemz\u00e9seket v\u00e9gezzen, p\u00e9ld\u00e1ul felismerje a nevezetess\u00e9geket, t\u00f6rt\u00e9nelmi kontextust biztos\u00edtson vagy seg\u00edtsen a navig\u00e1ci\u00f3ban. A technol\u00f3gia \u00edg\u00e9retesnek t\u0171nik olyan ter\u00fcleteken is, mint a v\u00e1rosfejleszt\u00e9s, a k\u00f6rnyezetv\u00e9delem, valamint a keres\u00e9si \u00e9s ment\u00e9si m\u0171veletek.<\/p>\n\n<h2 class=\"wp-block-heading\">A k\u00e9pfelismer\u00e9s j\u00f6v\u0151je<\/h2>\n\n<p>A PlaNethez hasonl\u00f3 neur\u00e1lis h\u00e1l\u00f3zatok jelent\u0151s el\u0151rel\u00e9p\u00e9st jelentenek a k\u00e9pelemz\u00e9sben. A kutat\u00f3k olyan j\u00f6v\u0151t k\u00e9pzelnek el, amelyben ezek a rendszerek m\u00e9g kifinomultabb\u00e1 v\u00e1lnak, lehet\u0151v\u00e9 t\u00e9ve sz\u00e1mukra, hogy tanuljanak egym\u00e1st\u00f3l, \u00e9s egyre \u00f6sszetettebb feladatokat hajtsanak v\u00e9gre. Ahogy az MI folyamatosan fejl\u0151dik, tov\u00e1bbi \u00e1tt\u00f6r\u00e9seket v\u00e1rhatunk, amelyek jav\u00edtj\u00e1k a vizu\u00e1lis vil\u00e1g meg\u00e9rt\u00e9s\u00e9nek \u00e9s azzal val\u00f3 interakci\u00f3 k\u00e9pess\u00e9g\u00e9t.<\/p>\n\n<h2 class=\"wp-block-heading\">Kieg\u00e9sz\u00edt\u0151 inform\u00e1ci\u00f3k<\/h2>\n\n<ul class=\"wp-block-list\">\n<li>A PlaNet pontoss\u00e1g\u00e1t hatalmas k\u00e9pz\u00e9si adatk\u00e9szlet\u00e9nek \u00e9s fejlett g\u00e9pi tanul\u00e1si algoritmusainak k\u00f6sz\u00f6nheti.<\/li>\n<li>A PlaNet lehets\u00e9ges alkalmaz\u00e1sai t\u00falmutatnak a k\u00e9pfelismer\u00e9sen, ide\u00e9rtve a t\u00e1rgyfelismer\u00e9st, az arcfelismer\u00e9st \u00e9s az orvosi k\u00e9pelemz\u00e9st.<\/li>\n<li>Ahogy a neur\u00e1lis h\u00e1l\u00f3zatok egyre er\u0151sebbek lesznek, a k\u00e9pfelismer\u00e9s pontoss\u00e1ga \u00e9s hat\u00f3k\u00f6re is tov\u00e1bb fog fejl\u0151dni.<\/li>\n<li>Figyelembe kell venni az MI-alap\u00fa k\u00e9pfelismer\u00e9s etikai vonatkoz\u00e1sait, k\u00fcl\u00f6n\u00f6sen az adatv\u00e9delem \u00e9s a megfigyel\u00e9s tekintet\u00e9ben.<\/li>\n<\/ul>","protected":false},"excerpt":{"rendered":"<p>A Google \u00faj mesters\u00e9ges intelligenci\u00e1ja pontosan meghat\u00e1rozza a fot\u00f3k hely\u00e9t \u00cdme a PlaNet: a Google k\u00e9pfelismer\u0151 neur\u00e1lis h\u00e1l\u00f3zata A mesters\u00e9ges intelligencia (MI) ter\u00e9n a Google jelent\u0151s el\u0151rel\u00e9p\u00e9st tett a PlaNet&hellip;<\/p>\n","protected":false},"author":6,"featured_media":23799,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2224],"tags":[1541,19771,19772,7721,1254,3506,19773,18056],"class_list":["post-14756","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-artificial-intelligence","tag-machine-learning","tag-image-analysis","tag-image-localization","tag-deep-learning","tag-artificial-intelligence","tag-neural-networks","tag-planet","tag-computer-vision"],"_links":{"self":[{"href":"https:\/\/www.lifescienceart.com\/hu\/wp-json\/wp\/v2\/posts\/14756","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.lifescienceart.com\/hu\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.lifescienceart.com\/hu\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.lifescienceart.com\/hu\/wp-json\/wp\/v2\/users\/6"}],"replies":[{"embeddable":true,"href":"https:\/\/www.lifescienceart.com\/hu\/wp-json\/wp\/v2\/comments?post=14756"}],"version-history":[{"count":1,"href":"https:\/\/www.lifescienceart.com\/hu\/wp-json\/wp\/v2\/posts\/14756\/revisions"}],"predecessor-version":[{"id":14757,"href":"https:\/\/www.lifescienceart.com\/hu\/wp-json\/wp\/v2\/posts\/14756\/revisions\/14757"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.lifescienceart.com\/hu\/wp-json\/wp\/v2\/media\/23799"}],"wp:attachment":[{"href":"https:\/\/www.lifescienceart.com\/hu\/wp-json\/wp\/v2\/media?parent=14756"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.lifescienceart.com\/hu\/wp-json\/wp\/v2\/categories?post=14756"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.lifescienceart.com\/hu\/wp-json\/wp\/v2\/tags?post=14756"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}