{"id":35473,"date":"2019-10-31T22:04:32","date_gmt":"2019-10-31T19:04:32","guid":{"rendered":"https:\/\/prohoster.info\/blog\/novaya-statya-vychislitelnaya-fotografiya\/"},"modified":"2019-10-31T22:04:32","modified_gmt":"2019-10-31T19:04:32","slug":"novaya-statya-vychislitelnaya-fotografiya","status":"publish","type":"post","link":"https:\/\/prohoster.info\/ro\/blog\/news\/novaya-statya-vychislitelnaya-fotografiya","title":{"rendered":"Articol nou: Fotografia Computa\u021bional\u0103","gt_translate_keys":[{"key":"rendered","format":"text"}]},"content":{"rendered":"<p>Originalul articolului a fost publicat pe site <noindex><a rel=\"nofollow noopener noreferrer\" href=\"https:\/\/vas3k.ru\/\" target=\"_blank\">Vastrik.ro<\/a><\/noindex>&nbsp;\u0219i publicat pe 3DNews cu permisiunea autorului. Prezent\u0103m textul complet al articolului, cu excep\u021bia num\u0103rului mare de linkuri \u2013 acestea vor fi utile celor care sunt cu adev\u0103rat interesa\u021bi de subiect \u0219i doresc s\u0103 studieze aspectele teoretice ale fotografiei computa\u021bionale mai \u00een profunzime, dar pentru publicul larg, am considerat c\u0103 acest material este excesiv.&nbsp;&nbsp;<\/p>\n<p>Ast\u0103zi, nicio prezentare de smartphone nu este complet\u0103 f\u0103r\u0103 a analiza camera sa. \u00cen fiecare lun\u0103 auzim despre un nou succes al camerelor mobile: Google \u00eenva\u021b\u0103 Pixel s\u0103 fotografieze \u00een \u00eentuneric, Huawei zoom-eaz\u0103 ca un binoclu, Samsung adaug\u0103 lidar, iar Apple creeaz\u0103 cele mai rotunde col\u021buri din lume. Inova\u021biile curg abundent de pretutindeni.<\/p>\n<p>\u00cen schimb, aparatele foto DSLR p\u0103reau c\u0103 stau pe loc. Sony ne surprinde anual cu noi senzori, iar produc\u0103torii \u00ee\u0219i actualizeaz\u0103 cu lejeritate ultima cifr\u0103 din versiune, continu\u00e2nd s\u0103 fumeze relaxa\u021bi \u00een col\u021b. Am o camer\u0103 DSLR pe birou cost\u00e2nd $3000, dar \u00een c\u0103l\u0103torii \u00eemi iau iPhone-ul. De ce?<\/p>\n<p>A\u0219a cum spunea un clasic \u2013 am ie\u0219it pe internet cu aceast\u0103 \u00eentrebare. Acolo discut\u0103 despre diverse \u201ealgoritmi\u201d \u0219i \u201ere\u021bele neuronale\u201d, f\u0103r\u0103 s\u0103 aib\u0103 idee cum anume acestea afecteaz\u0103 fotografia. Jurnali\u0219tii rostogolesc cu voce tare num\u0103rul de megapixeli, bloggerii se filmeaz\u0103 \u00een grup cu unboxing-uri pl\u0103tite, iar esteti se bucur\u0103 de \u201epercep\u021bia senzorial\u0103 a paletei de culori a senzorului\u201d. Totul ca de obicei.<\/p>\n<p>A trebuit s\u0103 m\u0103 a\u0219ez, s\u0103-mi pierd jum\u0103tate din via\u021b\u0103 \u0219i s\u0103 \u00een\u021beleg totul pe cont propriu. \u00cen acest articol voi povesti ce am aflat.<\/p>\n<h2><noindex><a rel=\"nofollow\" href=\"#contents\" class=\"right paragraph tooltip\" data-tooltip=\"\u0412\u0435\u0440\u043d\u0443\u0442\u044c\u0441\u044f \u043a \u043e\u0433\u043b\u0430\u0432\u043b\u0435\u043d\u0438\u044e\">\ud83d\udd3d<\/a><\/noindex><noindex><a rel=\"nofollow\" name=\"\u0427\u0442\u043e \u0442\u0430\u043a\u043e\u0435 \u0432\u044b\u0447\u0438\u0441\u043b\u0438\u0442\u0435\u043b\u044c\u043d\u0430\u044f \u0444\u043e\u0442\u043e\u0433\u0440\u0430\u0444\u0438\u044f?\" href=\"#\u0427\u0442\u043e \u0442\u0430\u043a\u043e\u0435 \u0432\u044b\u0447\u0438\u0441\u043b\u0438\u0442\u0435\u043b\u044c\u043d\u0430\u044f \u0444\u043e\u0442\u043e\u0433\u0440\u0430\u0444\u0438\u044f?\" class=\"right paragraph tooltip\" data-tooltip=\"\u0421\u0441\u044b\u043b\u043a\u0430 \u043d\u0430 \u044d\u0442\u0443 \u0433\u043b\u0430\u0432\u0443\">#<\/a><\/noindex>Ce este fotografia computa\u021bional\u0103?<\/h2>\n<p>Peste tot, inclusiv pe Wikipedia, se ofer\u0103 o defini\u021bie asem\u0103n\u0103toare: fotografia computa\u021bional\u0103 \u2013 orice tehnici de captur\u0103 \u0219i procesare a imaginilor, unde \u00een locul transform\u0103rilor optice se folosesc calcule digitale. Aici totul este bine, cu excep\u021bia faptului c\u0103 nu explic\u0103 nimic. Se potrive\u0219te chiar \u0219i cu autofocusul, dar nu include op\u021biunile de pelicul\u0103, care deja ne-au adus multe beneficii. Ambiguitatea defini\u021biilor oficiale sugereaz\u0103 c\u0103 nu avem idee despre ce vorbim.<\/p>\n<p>Pionierul fotografiei computa\u021bionale, profesorul de la Stanford Marc Levoy (care acum r\u0103spunde de camera din Google Pixel) ofer\u0103 o alt\u0103 defini\u021bie \u2013 un set de metode de vizualizare computerizat\u0103 care \u00eembun\u0103t\u0103\u021besc sau extind capacit\u0103\u021bile fotografiei digitale, prin utilizarea c\u0103rora se ob\u021bine o fotografie obi\u0219nuit\u0103 care nu ar putea fi realizat\u0103 tehnic cu aceast\u0103 camer\u0103 \u00een mod tradi\u021bional. \u00cen acest articol, m\u0103 voi baza exact pe defini\u021bia lui.<\/p>\n<p>A\u0219adar, smartphone-urile au fost de vin\u0103.<\/p>\n<p style=\"text-align: center;\">Smartphone-urile nu aveau alt\u0103 op\u021biune dec\u00e2t s\u0103 dea via\u021b\u0103 unei noi forme de fotografie \u2013 fotografia computa\u021bional\u0103.<\/p>\n<p>Micu\u021bele lor matrice zgomotoase \u0219i obiectivele minusculi cu o capacitate redus\u0103 de iluminare, prin toate legile fizicii ar fi trebuit s\u0103 aduc\u0103 doar durere \u0219i suferin\u021b\u0103. \u0218i au adus, p\u00e2n\u0103 c\u00e2nd dezvoltatorii lor nu s-au g\u00e2ndit s\u0103 foloseasc\u0103 ingenios punctele forte pentru a compensa sl\u0103biciunile \u2013 obturatoare electronice rapide, procesoare puternice \u0219i software.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/9735433adae34a087804b604eff125d4.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Cele mai multe cercet\u0103ri sonore \u00een domeniul fotografiei computa\u021bionale au avut loc \u00eentre anii 2005-2015, ceea ce \u00een \u0219tiin\u021b\u0103 este considerat practic ieri. Chiar acum, sub ochii no\u0219tri \u0219i \u00een buzunarele noastre, se dezvolt\u0103 un nou domeniu de cuno\u0219tin\u021be \u0219i tehnologii care nu a existat \u00eenainte.<\/p>\n<p>Fotografia computa\u021bional\u0103 nu se rezum\u0103 doar la selfie-uri cu bokeh neuro. Fotografia recent\u0103 a unei g\u0103uri negre nu ar fi ap\u0103rut f\u0103r\u0103 metodele de fotografie computa\u021bional\u0103. Pentru a captura o astfel de imagine cu un telescop obi\u0219nuit, ar trebui s\u0103-l facem de dimensiunea P\u0103m\u00e2ntului. Totu\u0219i, combin\u00e2nd datele de la opt radiotelescopuri din diferite col\u021buri ale globului nostru \u0219i scriind c\u00e2teva scripturi \u00een Python, am ob\u021binut prima fotografie din lume a orizontului evenimentelor. Potrivit \u0219i pentru selfie-uri.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/696e1ec7834e8bf47bff2aa8965b903a.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<h2><noindex><a rel=\"nofollow\" href=\"#contents\" class=\"right paragraph tooltip\" data-tooltip=\"\u0412\u0435\u0440\u043d\u0443\u0442\u044c\u0441\u044f \u043a \u043e\u0433\u043b\u0430\u0432\u043b\u0435\u043d\u0438\u044e\">\ud83d\udd3d<\/a><\/noindex><noindex><a rel=\"nofollow\" name=\"\u041d\u0430\u0447\u0430\u043b\u043e: \u0446\u0438\u0444\u0440\u043e\u0432\u0430\u044f \u043e\u0431\u0440\u0430\u0431\u043e\u0442\u043a\u0430\" href=\"#\u041d\u0430\u0447\u0430\u043b\u043e: \u0446\u0438\u0444\u0440\u043e\u0432\u0430\u044f \u043e\u0431\u0440\u0430\u0431\u043e\u0442\u043a\u0430\" class=\"right paragraph tooltip\" data-tooltip=\"\u0421\u0441\u044b\u043b\u043a\u0430 \u043d\u0430 \u044d\u0442\u0443 \u0433\u043b\u0430\u0432\u0443\">#<\/a><\/noindex>\u00cenceput: procesarea digital\u0103<\/h2>\n<p>S\u0103 ne imagin\u0103m c\u0103 ne-am \u00eentors \u00een 2007. Mama noastr\u0103 este anarhie, iar fotografiile noastre sunt zgomotoase jpeg-uri de 0,6 MP, realizate cu skateboarding. Aproape atunci ne-a ap\u0103rut prima dorin\u021b\u0103 irezistibil\u0103 de a aplica preset\u0103ri pentru a ascunde sl\u0103biciunile matricelor mobile. S\u0103 nu ne priv\u0103m de pl\u0103cere.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/4078f9ff31730ac7b4dec18d4408a707.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<h2><noindex><a rel=\"nofollow\" href=\"#contents\" class=\"right paragraph tooltip\" data-tooltip=\"\u0412\u0435\u0440\u043d\u0443\u0442\u044c\u0441\u044f \u043a \u043e\u0433\u043b\u0430\u0432\u043b\u0435\u043d\u0438\u044e\">\ud83d\udd3d<\/a><\/noindex><noindex><a rel=\"nofollow\" name=\"\u041c\u0430\u0442\u0430\u043d \u0438 \u0438\u043d\u0441\u0442\u0430\u0433\u0440\u0430\u043c\" href=\"#\u041c\u0430\u0442\u0430\u043d \u0438 \u0438\u043d\u0441\u0442\u0430\u0433\u0440\u0430\u043c\" class=\"right paragraph tooltip\" data-tooltip=\"\u0421\u0441\u044b\u043b\u043a\u0430 \u043d\u0430 \u044d\u0442\u0443 \u0433\u043b\u0430\u0432\u0443\">#<\/a><\/noindex>Matematica \u0219i Instagram<\/h2>\n<p>Odat\u0103 cu lansarea Instagramului, toat\u0103 lumea a devenit obsedat\u0103 de filtre. Ca persoan\u0103 care, la un moment dat, a efectuat reverse engineering pe X-Pro II, Lo-Fi \u0219i Valencia \u00een scopuri, desigur, de cercetare (kek), \u00eemi amintesc c\u0103 acestea constau din trei componente:<\/p>\n<ul>\n<li>Set\u0103rile de culoare (Nuanta, Saturatie, Luminozitate, Contrast, Niveluri etc.) sunt coeficien\u021bi digitali simpli, exact ca cele din preset\u0103rile folosite de fotografi din cele mai vechi timpuri.<\/li>\n<li>H\u0103r\u021bile de tonalitate (Tone Mapping) sunt vectori de valori, fiecare dintre ele spun\u00e2ndu-ne: \u201eCuloarea ro\u0219ie cu nuan\u021ba 128 trebuie transformat\u0103 \u00een nuan\u021ba 240\u201d.<\/li>\n<li>Overlay-uri \u2014 imagini semi-transparente cu praf, granulare, vinietare \u0219i tot ce poate fi suprapus pentru a ob\u021bine un efect deloc banal de film vechi. Nu a fost \u00eentotdeauna prezent.&nbsp; &nbsp;<\/li>\n<\/ul>\n<p>Filtrele moderne nu s-au \u00eendep\u0103rtat mult de aceast\u0103 triad\u0103, ci au devenit doar pu\u021bin mai complexe din punct de vedere matematic. Odat\u0103 cu apari\u021bia shader-elor hardware \u0219i OpenCL pe smartphone-uri, acestea au fost rapid adaptate pentru GPU, ceea ce era considerat extrem de tare. Pentru anul 2012, desigur. Ast\u0103zi, orice elev poate crea ceva similar folosind CSS, iar la absolvire nu va primi nici m\u0103car un premiu.<\/p>\n<p>\u00cens\u0103 progresul filtrelor nu s-a oprit ast\u0103zi. De exemplu, echipa de la Dehancer se concentreaz\u0103 mult pe filtre non-liniare \u2014 \u00een loc de banalul ton-mapping, ei folosesc transform\u0103ri non-liniare mai complexe, ceea ce, potrivit lor, deschide mult mai multe oportunit\u0103\u021bi.<\/p>\n<p>Transform\u0103rile non-liniare pot genera o mul\u021bime de lucruri, dar sunt incredibil de complexe, iar noi, ca oameni, suntem incredibil de pro\u0219ti. Atunci c\u00e2nd \u0219tiin\u021ba ajunge la transform\u0103ri non-liniare, prefer\u0103m s\u0103 ne \u00eendrept\u0103m spre metode numerice \u0219i s\u0103 \u00eencorpor\u0103m inteligen\u021ba artificial\u0103, l\u0103s\u00e2nd-o s\u0103 scrie capodopere \u00een locul nostru. A\u0219a a fost \u0219i \u00een acest caz.<\/p>\n<h2><noindex><a rel=\"nofollow\" href=\"#contents\" class=\"right paragraph tooltip\" data-tooltip=\"\u0412\u0435\u0440\u043d\u0443\u0442\u044c\u0441\u044f \u043a \u043e\u0433\u043b\u0430\u0432\u043b\u0435\u043d\u0438\u044e\">\ud83d\udd3d<\/a><\/noindex><noindex><a rel=\"nofollow\" name=\"\u0410\u0432\u0442\u043e\u043c\u0430\u0442\u0438\u043a\u0430 \u0438 \u043c\u0435\u0447\u0442\u044b \u043e \u043a\u043d\u043e\u043f\u043a\u0435 &laquo;\u0448\u0435\u0434\u0435\u0432\u0440&raquo;\" href=\"#\u0410\u0432\u0442\u043e\u043c\u0430\u0442\u0438\u043a\u0430 \u0438 \u043c\u0435\u0447\u0442\u044b \u043e \u043a\u043d\u043e\u043f\u043a\u0435 &laquo;\u0448\u0435\u0434\u0435\u0432\u0440&raquo;\" class=\"right paragraph tooltip\" data-tooltip=\"\u0421\u0441\u044b\u043b\u043a\u0430 \u043d\u0430 \u044d\u0442\u0443 \u0433\u043b\u0430\u0432\u0443\">#<\/a><\/noindex>Automatizare \u0219i vise despre butonul \u201ecapodoper\u0103\u201d<\/h2>\n<p>C\u00e2nd toat\u0103 lumea s-a obi\u0219nuit cu filtrele, am \u00eenceput s\u0103 le integr\u0103m direct \u00een camere. Istoria ascunde cine anume a fost primul produc\u0103tor, dar pentru a \u00een\u021belege c\u00e2t de demult s-a \u00eent\u00e2mplat \u2014 \u00een iOS 5.0, care a fost lansat \u00een 2011, exista deja un API public pentru \u00cembun\u0103t\u0103\u021birea Automat\u0103 a Imaginilor. Numai lui Jobs \u00eei este cunoscut c\u00e2t de mult a fost folosit \u00eenainte de a fi deschis publicului.<\/p>\n<p>Automatica f\u0103cea acelea\u0219i lucruri ca \u0219i fiecare dintre noi, deschiz\u00e2nd o fotografie \u00een editor, \u2013 ajusta umbrele \u0219i luminozitatea, ad\u0103uga saturarea, elimina ochii ro\u0219ii \u0219i fixa culoarea fe\u021bei. Utilizatorii nu aveau nici cea mai mic\u0103 idee c\u0103 \u201ecamera dramatic \u00eembun\u0103t\u0103\u021bit\u0103\u201d din noul smartphone era rezultatul a c\u00e2torva noi \u0219ed\u0103re. P\u00e2n\u0103 la lansarea Google Pixel \u0219i \u00eenceputul hype-ului despre fotografia computa\u021bional\u0103, mai erau \u00eenc\u0103 cinci ani.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/480b0c337a37ce33a178794104c2d541.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Ast\u0103zi, b\u0103t\u0103liile pentru butonul \u201ecapodoper\u0103\u201d s-au mutat \u00een domeniul \u00eenv\u0103\u021b\u0103rii automate. Dup\u0103 ce s-au jucat cu ton-mapping-ul, to\u021bi au \u00eenceput s\u0103 antreneze CNN-uri \u0219i GAN-uri pentru a muta butoanele \u00een locul utilizatorului. Cu alte cuvinte, pe baza imaginii de intrare, se stabile\u0219te un set de parametrii optimi care apropie imaginea de o anumit\u0103 \u00een\u021belegere subiectiv\u0103 a \u201eunei fotografii bune\u201d. Implementat \u00een acela\u0219i Pixelmator Pro \u0219i \u00een alte editoare. Func\u021bioneaz\u0103, cum se poate ghici, nu foarte bine \u0219i nu \u00een toate cazurile.&nbsp;<\/p>\n<h2><noindex><a rel=\"nofollow\" href=\"#contents\" class=\"right paragraph tooltip\" data-tooltip=\"\u0412\u0435\u0440\u043d\u0443\u0442\u044c\u0441\u044f \u043a \u043e\u0433\u043b\u0430\u0432\u043b\u0435\u043d\u0438\u044e\">\ud83d\udd3d<\/a><\/noindex><noindex><a rel=\"nofollow\" name=\"\u0421\u0442\u0435\u043a\u0438\u043d\u0433 &mdash; 90% \u0443\u0441\u043f\u0435\u0445\u0430 \u043c\u043e\u0431\u0438\u043b\u044c\u043d\u044b\u0445 \u043a\u0430\u043c\u0435\u0440\" href=\"#\u0421\u0442\u0435\u043a\u0438\u043d\u0433 &mdash; 90% \u0443\u0441\u043f\u0435\u0445\u0430 \u043c\u043e\u0431\u0438\u043b\u044c\u043d\u044b\u0445 \u043a\u0430\u043c\u0435\u0440\" class=\"right paragraph tooltip\" data-tooltip=\"\u0421\u0441\u044b\u043b\u043a\u0430 \u043d\u0430 \u044d\u0442\u0443 \u0433\u043b\u0430\u0432\u0443\">#<\/a><\/noindex>Stacking \u2013 90% din succesul camerelor mobile<\/h2>\n<p>Adev\u0103rata fotografie computa\u021bional\u0103 a \u00eenceput cu stacking-ul \u2013 suprapunerea mai multor fotografii una peste alta. Pentru un smartphone nu este o problem\u0103 s\u0103 captureze o duzin\u0103 de cadre \u00een jum\u0103tate de secund\u0103. \u00cen camerele lor nu exist\u0103 p\u0103r\u021bi mecanice lente: diafragma este fix\u0103, iar \u00een loc de oblonul mobil exist\u0103 un obturator electronic. Procesorul d\u0103 pur \u0219i simplu un ordin matricei, spun\u00e2ndu-i c\u00e2te microsecunde s\u0103 capteze fotonii s\u0103lbatici, iar el citeste rezultatul.<\/p>\n<p>Tehnic, telefonul poate face fotografii cu viteza unui video, iar video cu rezolu\u021bia unei fotografii, dar totul depinde de viteza magistralei \u0219i a procesorului. De aceea, se impun \u00eentotdeauna limite software.<\/p>\n<p>Stacking-ul a fost cu noi de mult timp. Chiar \u0219i bunicii no\u0219tri puneau plugin-uri \u00een Photoshop 7.0 pentru a aduna mai multe fotografii \u00eentr-un HDR strident sau pentru a lipi o panoram\u0103 de 18000 \u00d7 600 pixeli \u0219i... de fapt, nimeni nu a reu\u0219it s\u0103 \u00een\u021beleag\u0103 ce s\u0103 fac\u0103 cu ele mai departe. Au fost vremuri bogate, din p\u0103cate, s\u0103lbatice.<\/p>\n<p>Acum am crescut \u0219i numim asta \u201efotografia epsilon\u201d \u2013 c\u00e2nd, modific\u00e2nd unul dintre parametrii camerei (expunerea, focalizarea, pozi\u021bia) \u0219i lipind cadrele ob\u021binute, ob\u021binem ceva ce nu putea fi capturat \u00eentr-un singur cadru. Dar acesta este un termen pentru teoreticieni, \u00een practic\u0103 s-a adaptat o alt\u0103 denumire \u2013 stacking. Ast\u0103zi, de fapt, pe el se bazeaz\u0103 90% din toate inova\u021biile din camerele mobile.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/f3a6fd9c8260116a371eeb03bbf61536.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Un lucru la care mul\u021bi nu se g\u00e2ndesc, dar care este important pentru \u00een\u021belegerea \u00eentregii fotografii mobile \u0219i computa\u021bionale: camera din smartphone-ul modern \u00eencepe s\u0103 fac\u0103 fotografii imediat ce deschide\u021bi aplica\u021bia ei. Ceea ce este logic, deoarece trebuie s\u0103 transmit\u0103 cumva imaginea pe ecran. Totu\u0219i, pe l\u00e2ng\u0103 ecran, ea salveaz\u0103 cadre de \u00eenalt\u0103 rezolu\u021bie \u00een propriul s\u0103u buffer circular, unde le p\u0103streaz\u0103 \u00eenc\u0103 c\u00e2teva secunde.<\/p>\n<p style=\"text-align: center;\">C\u00e2nd ap\u0103sa\u021bi butonul \u201ef\u0103 fotografie\u201d \u2013 de fapt, deja a fost realizat\u0103, camera pur \u0219i simplu ia ultima fotografie din buffer.<\/p>\n<p>Ast\u0103zi, orice camer\u0103 mobil\u0103 func\u021bioneaz\u0103 a\u0219a. Cel pu\u021bin, \u00een toate flagship-urile, nu \u00een cele din gunoaie. Bufferizarea permite nu doar realizarea unei \u00eent\u00e2rzieri zero a obturatorului, despre care fotografii viseaz\u0103 de mult, ci chiar \u0219i o \u00eent\u00e2rziere negativ\u0103; atunci c\u00e2nd ap\u0103sa\u021bi pe buton, smartphone-ul prive\u0219te \u00een trecut, descarc\u0103 ultimele 5-10 fotografii din buffer \u0219i \u00eencepe s\u0103 le analizeze frenetic \u0219i s\u0103 le combine. Nu mai trebuie s\u0103 a\u0219tepta\u021bi ca telefonul s\u0103 fac\u0103 capturi pentru HDR sau modul nocturn; pur \u0219i simplu le lua\u021bi din buffer, utilizatorul nici m\u0103car nu va observa.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/df4d90e674db51c1e8b43518300270da.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Apropo, exact prin intermediul \u00eent\u00e2rzierii negative a obturatorului s-a realizat Live Photo pe iPhone-uri, iar la HTC ceva similar exista \u00eenc\u0103 din 2013 sub denumirea ciudat\u0103 Zoe.<\/p>\n<h2><noindex><a rel=\"nofollow\" href=\"#contents\" class=\"right paragraph tooltip\" data-tooltip=\"\u0412\u0435\u0440\u043d\u0443\u0442\u044c\u0441\u044f \u043a \u043e\u0433\u043b\u0430\u0432\u043b\u0435\u043d\u0438\u044e\">\ud83d\udd3d<\/a><\/noindex><noindex><a rel=\"nofollow\" name=\"\u0421\u0442\u0435\u043a\u0438\u043d\u0433 \u043f\u043e \u044d\u043a\u0441\u043f\u043e\u0437\u0438\u0446\u0438\u0438 &mdash; HDR \u0438 \u0431\u043e\u0440\u044c\u0431\u0430 \u0441 \u043f\u0435\u0440\u0435\u043f\u0430\u0434\u0430\u043c\u0438 \u044f\u0440\u043a\u043e\u0441\u0442\u0438\" href=\"#\u0421\u0442\u0435\u043a\u0438\u043d\u0433 \u043f\u043e \u044d\u043a\u0441\u043f\u043e\u0437\u0438\u0446\u0438\u0438 &mdash; HDR \u0438 \u0431\u043e\u0440\u044c\u0431\u0430 \u0441 \u043f\u0435\u0440\u0435\u043f\u0430\u0434\u0430\u043c\u0438 \u044f\u0440\u043a\u043e\u0441\u0442\u0438\" class=\"right paragraph tooltip\" data-tooltip=\"\u0421\u0441\u044b\u043b\u043a\u0430 \u043d\u0430 \u044d\u0442\u0443 \u0433\u043b\u0430\u0432\u0443\">#<\/a><\/noindex>Stivuirea pe expunere \u2013 HDR \u0219i combaterea varia\u021biilor de luminozitate<\/h2>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/f05831e57c87ec1177cdef0dc578b46c.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Sunt capabile matricile aparatelor foto s\u0103 \u00eenregistreze toat\u0103 gama de luminozitate disponibil\u0103 ochiului nostru \u2013 este un subiect vechi de dispute. Unii spun c\u0103 nu, deoarece ochiul poate percepe p\u00e2n\u0103 la 25 de stopuri f, \u00een timp ce chiar \u0219i dintr-o matrice full-frame de top se pot extrage maximum 14. Al\u021bii consider\u0103 c\u0103 compara\u021bia este incorect\u0103, deoarece ochiul este ajutat de creier, care ajusteaz\u0103 automat pupila \u0219i completeaz\u0103 imaginea cu re\u021belele sale neuronale, iar intervalul dinamic imediat al ochiului nu este, de fapt, mai mare de 10-14 stopuri f. S\u0103 l\u0103s\u0103m aceste dispute celor mai buni g\u00e2nditori de pe internet.<\/p>\n<p>Faptul r\u0103m\u00e2ne: film\u00e2nd prietenii pe fundalul unui cer luminos f\u0103r\u0103 HDR pe orice camer\u0103 mobil\u0103, ob\u021bine\u021bi fie un cer normal cu fe\u021bele prietenilor negre, fie prieteni bine contura\u021bi, dar cu cerul ars de soare.<\/p>\n<p>Solu\u021bia a fost g\u00e2ndit\u0103 demult \u2013 extinderea intervalului de luminozitate prin HDR (High Dynamic Range). Este necesar s\u0103 se fac\u0103 c\u00e2teva cadre cu expuneri diferite \u0219i s\u0103 le combin\u0103m. Unul s\u0103 fie \u201enormal\u201d, altul mai luminos, al treilea mai \u00eentunecat. Lu\u0103m locurile \u00eentunecate din cadru luminos, umplem supraexpunerile cu cel \u00eentunecat \u2013 profit. R\u0103m\u00e2ne doar s\u0103 rezolv\u0103m problema bracketing-ului automat \u2013 c\u00e2t de mult s\u0103 deplas\u0103m expunerea fiec\u0103rui cadru, astfel \u00eenc\u00e2t s\u0103 nu exager\u0103m, dar stabilirea luminozit\u0103\u021bii medii a imaginii poate fi realizat\u0103 acum cu u\u0219urin\u021b\u0103 chiar \u0219i de un student \u00een anul doi de la o universitate tehnic\u0103.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/de7ba87bb60f8da89c650220d431248b.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Pe ultimele iPhone-uri, Pixel \u0219i Galaxy, modul HDR se activeaz\u0103 automat atunci c\u00e2nd un algoritm simplu din interiorul camerei determin\u0103 c\u0103 fotografia\u021bi ceva contrastant \u00eentr-o zi \u00eensorit\u0103. Pute\u021bi observa chiar cum telefonul comut\u0103 modul de \u00eenregistrare \u00een buffer pentru a salva cadrele deplasate pe expunere; \u00een camer\u0103, fps-ul scade, iar imaginea devine mai vibrant\u0103. Momentul comut\u0103rii este foarte vizibil pe iPhone-ul meu X, dac\u0103 fotografiez afar\u0103. Uita\u021bi-v\u0103 la smartphone-ul dumneavoastr\u0103 \u0219i data viitoare.<\/p>\n<p>Un dezavantaj al HDR cu bracketing pe expunere \u2013 incapacitatea sa de a func\u021biona \u00een condi\u021bii de iluminare slab\u0103. Chiar \u0219i cu lumina unei l\u0103mpi de camer\u0103, imaginile devin at\u00e2t de \u00eentunecate \u00eenc\u00e2t computerul nu le poate alinia \u0219i lipi. Pentru a rezolva problema luminilor, \u00een 2013 Google a prezentat o alt\u0103 abordare pentru HDR \u00een smartphone-ul Nexus lansat atunci. Acesta a folosit stacked temporar.<\/p>\n<h2><noindex><a rel=\"nofollow\" href=\"#contents\" class=\"right paragraph tooltip\" data-tooltip=\"\u0412\u0435\u0440\u043d\u0443\u0442\u044c\u0441\u044f \u043a \u043e\u0433\u043b\u0430\u0432\u043b\u0435\u043d\u0438\u044e\">\ud83d\udd3d<\/a><\/noindex><noindex><a rel=\"nofollow\" name=\"\u0421\u0442\u0435\u043a\u0438\u043d\u0433 \u043f\u043e \u0432\u0440\u0435\u043c\u0435\u043d\u0438 &mdash; \u0441\u0438\u043c\u0443\u043b\u044f\u0446\u0438\u044f \u0434\u043b\u0438\u043d\u043d\u043e\u0439 \u0432\u044b\u0434\u0435\u0440\u0436\u043a\u0438 \u0438 \u0442\u0430\u0439\u043c\u043b\u0430\u043f\u0441\" href=\"#\u0421\u0442\u0435\u043a\u0438\u043d\u0433 \u043f\u043e \u0432\u0440\u0435\u043c\u0435\u043d\u0438 &mdash; \u0441\u0438\u043c\u0443\u043b\u044f\u0446\u0438\u044f \u0434\u043b\u0438\u043d\u043d\u043e\u0439 \u0432\u044b\u0434\u0435\u0440\u0436\u043a\u0438 \u0438 \u0442\u0430\u0439\u043c\u043b\u0430\u043f\u0441\" class=\"right paragraph tooltip\" data-tooltip=\"\u0421\u0441\u044b\u043b\u043a\u0430 \u043d\u0430 \u044d\u0442\u0443 \u0433\u043b\u0430\u0432\u0443\">#<\/a><\/noindex>Stivuirea temporal\u0103 \u2013 simularea unei expuneri lungi \u0219i a unui timelapse<\/h2>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/4fe1ab5f044e2097f11cf582926c2713.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Stacking temporal permite ob\u021binerea unei expuneri lungi printr-o serie de expuneri scurte. Primii care au experimentat au fost pasiona\u021bii care filmau urme de stele pe cerul nop\u021bii, care g\u0103seau incomod s\u0103 \u021bin\u0103 obturatorul deschis timp de dou\u0103 ore. Erau greu de calculat toate set\u0103rile din timp iar cea mai mic\u0103 mi\u0219care distrugea \u00eentregul cadru. Ei au decis s\u0103 \u021bin\u0103 obturatorul deschis doar c\u00e2teva minute, dar de mai multe ori, iar apoi se \u00eentorceau acas\u0103 \u0219i combinau cadrele ob\u021binute \u00een Photoshop.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/3985acce3f0bd1e1ceab9364f4f98d91.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Astfel, camera nu fotografia niciodat\u0103 cu o expunere lung\u0103, dar noi ob\u021bineam efectul imita\u021biei ei prin combinarea mai multor cadre trase consecutiv. Exist\u0103 deja o mul\u021bime de aplica\u021bii pentru smartphone-uri care folosesc acest truc, dar toate acestea nu mai sunt necesare din moment ce func\u021bia a fost ad\u0103ugat\u0103 \u00een aproape toate camerele standard. Ast\u0103zi, chiar \u0219i un iPhone poate combina cu u\u0219urin\u021b\u0103 o expunere lung\u0103 din Live Photo.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/e6bf5859cab4f105990859735f6a48e7.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>S\u0103 ne \u00eentoarcem la Google cu HDR-ul s\u0103u nocturn. S-a dovedit c\u0103, prin bracketingul pe timp, po\u021bi realiza un HDR decent \u00een \u00eentuneric. Tehnologia a ap\u0103rut pentru prima dat\u0103 \u00een Nexus 5 \u0219i a fost numit\u0103 HDR+. Celelalte telefoane cu Android au primit-o ca un fel de cadou. Tehnologia este at\u00e2t de popular\u0103 \u00eenc\u00e2t este l\u0103udat\u0103 chiar \u0219i \u00een prezent\u0103rile celor mai recente Pixel.<\/p>\n<p>HDR+ func\u021bioneaz\u0103 destul de simplu: dac\u0103 detecteaz\u0103 c\u0103 fotografia\u021bi \u00een \u00eentuneric, camera descarc\u0103 8-15 dintre cele mai recente fotografii \u00een RAW din buffer pentru a le suprapune. Astfel, algoritmul colecteaz\u0103 mai multe informa\u021bii despre zonele \u00eentunecate ale imaginii pentru a minimiza zgomotul \u2013 pixeli \u00een care, din diverse motive, camera nu a reu\u0219it s\u0103 colecteze toat\u0103 informa\u021bia.<\/p>\n<p>Este ca \u0219i cum nu a\u021bi \u0219ti cum arat\u0103 o capibara \u0219i a\u021bi cere cinci persoane s\u0103 o descrie \u2013 pove\u0219tile lor ar fi aproximativ similare, dar fiecare ar men\u021biona un detaliu unic. Astfel, a\u021bi aduna mai multe informa\u021bii dec\u00e2t dac\u0103 a\u021bi \u00eentreba o singur\u0103 persoan\u0103. La fel este \u0219i cu pixelii.<\/p>\n<p>Combinarea imaginilor realizate din acela\u0219i punct ofer\u0103 acela\u0219i efect fals de expunere lung\u0103, ca \u0219i \u00een cazul stelelor de mai sus. Expunerea zecilor de cadre se adun\u0103, erorile dintr-unul sunt minimizate \u00een celelalte. Imagina\u021bi-v\u0103 c\u00e2t de des ar trebui s\u0103 ap\u0103sa\u021bi obturatorul DSLR-ului pentru a ob\u021bine a\u0219a ceva.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/eb7e3d8571ee4ae7eb7e62a47604ae1d.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>R\u0103m\u00e2nea doar s\u0103 rezolv\u0103m problema corect\u0103rii automate a culorilor \u2013 imaginile realizate \u00een \u00eentuneric sunt de obicei doar galbene sau verzi, iar noi, pare-se, vrem intensitatea luminii de zi. \u00cen versiunile anterioare ale HDR+ s-a apelat la ajustarea simpl\u0103 a set\u0103rilor, ca \u00een filtrele de tip Instagram. Apoi, s-au folosit re\u021bele neuronale.<\/p>\n<p>Astfel a ap\u0103rut Night Sight \u2013 tehnologia de 'fotografie nocturn\u0103' \u00een Pixel 2 \u0219i 3. \u00cen descriere se spune: 'Tehnici de \u00eenv\u0103\u021bare automat\u0103 construite pe baza HDR+, care fac ca Night Sight s\u0103 func\u021bioneze'. Practic, aceasta este automatizarea etapei de corectare a culorilor. Ma\u0219ina a fost antrenat\u0103 pe un set de date cu fotografii '\u00eenainte' \u0219i 'dup\u0103', pentru a transforma orice set de fotografii \u00eentunecate \u00eentr-o imagine frumoas\u0103.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/3451cbbeaa58a77f5ce9f88b7ac0de79.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Setul de date, de altfel, a fost publicat \u00een acces deschis. Poate c\u0103 b\u0103ie\u021bii de la Apple \u00eel vor lua \u0219i \u00een sf\u00e2r\u0219it \u00ee\u0219i vor \u00eenv\u0103\u021ba rachetele de sticl\u0103 s\u0103 fac\u0103 fotografii corecte \u00een \u00eentuneric.<\/p>\n<p>\u00cen plus, \u00een Night Sight este utilizat\u0103 calcularea vectorului de mi\u0219care al obiectelor din cadru pentru a normaliza neclarit\u0103\u021bile, care vor ap\u0103rea \u00eentotdeauna la expuneri lungi. Astfel, smartphone-ul poate lua p\u0103r\u021bile clare din alte cadre \u0219i le poate lipi.<\/p>\n<h2><noindex><a rel=\"nofollow\" href=\"#contents\" class=\"right paragraph tooltip\" data-tooltip=\"\u0412\u0435\u0440\u043d\u0443\u0442\u044c\u0441\u044f \u043a \u043e\u0433\u043b\u0430\u0432\u043b\u0435\u043d\u0438\u044e\">\ud83d\udd3d<\/a><\/noindex><noindex><a rel=\"nofollow\" name=\"\u0421\u0442\u0435\u043a\u0438\u043d\u0433 \u043f\u043e \u0434\u0432\u0438\u0436\u0435\u043d\u0438\u044e &mdash; \u043f\u0430\u043d\u043e\u0440\u0430\u043c\u0430, \u0441\u0443\u043f\u0435\u0440\u0437\u0443\u043c \u0438 \u0431\u043e\u0440\u044c\u0431\u0430 \u0441 \u0448\u0443\u043c\u0430\u043c\u0438\" href=\"#\u0421\u0442\u0435\u043a\u0438\u043d\u0433 \u043f\u043e \u0434\u0432\u0438\u0436\u0435\u043d\u0438\u044e &mdash; \u043f\u0430\u043d\u043e\u0440\u0430\u043c\u0430, \u0441\u0443\u043f\u0435\u0440\u0437\u0443\u043c \u0438 \u0431\u043e\u0440\u044c\u0431\u0430 \u0441 \u0448\u0443\u043c\u0430\u043c\u0438\" class=\"right paragraph tooltip\" data-tooltip=\"\u0421\u0441\u044b\u043b\u043a\u0430 \u043d\u0430 \u044d\u0442\u0443 \u0433\u043b\u0430\u0432\u0443\">#<\/a><\/noindex>Stivuirea pe mi\u0219care \u2013 panoram\u0103, super zoom \u0219i lupta cu zgomotul<\/h2>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/c03260ea8ad60358065ec81ed0539063.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Panorama \u2013 o distrac\u021bie popular\u0103 \u00een r\u00e2ndul locuitorilor din zonele rurale. P\u00e2n\u0103 acum nu se \u0219tie s\u0103 fi existat cazuri \u00een care o 'fotografie cu c\u00e2rna\u021bi' s\u0103 fie interesant\u0103 pentru altcineva dec\u00e2t pentru autorul s\u0103u, dar nu o putem omite \u2013 pentru mul\u021bi, aceasta a fost \u00eenceputul stivuiri.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/bf5955b2f32018ab90c12d6da41107c1.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Prima util\u0103 a panoramei este ob\u021binerea unei fotografii cu o rezolu\u021bie mai mare dec\u00e2t permite senzorul camerei prin unirea mai multor cadre. Fotograficii folosesc de mult diverse software-uri pentru a crea ceea ce se nume\u0219te fotografii cu superrezolu\u021bie, c\u00e2nd fotografiile u\u0219or deplasate se completeaz\u0103 \u00eentre ele \u00eentre pixeli. Astfel, se poate ob\u021bine o imagine de sute de gigapixeli, ceea ce este foarte util dac\u0103 vrei s\u0103 o tip\u0103re\u0219ti pe un panou publicitar de dimensiunea unei case.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/41f0d56913cd7b4c9e407345bf7383ce.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>O alt\u0103 abordare, deja mai interesant\u0103, este Pixel Shifting. Unele camere mirrorless precum Sony \u0219i Olympus au \u00eenceput s\u0103 o sus\u021bin\u0103 \u00eenc\u0103 din 2014, dar era necesar s\u0103 lipse\u0219ti rezultatul manual. Inova\u021bii tipice pentru camerele mari.<\/p>\n<p>Smartphone-urile au excelat aici dintr-un motiv amuzant: atunci c\u00e2nd faci o fotografie, m\u00e2inile tale tremur\u0103. Aceast\u0103 problem\u0103 aparent trivial\u0103 a stat la baza implement\u0103rii nativ\u0103 a superrezolu\u021biei pe smartphone-uri.<\/p>\n<p>Pentru a \u00een\u021belege cum func\u021bioneaz\u0103, trebuie s\u0103 ne amintim cum este construit senzorul oric\u0103rei camere foto. Fiecare pixel (fotodiod) este capabil s\u0103 fixeze doar intensitatea luminii, adic\u0103 num\u0103rul de fotoni care au intrat. Totu\u0219i, pixelul nu poate m\u0103sura culoarea acestora (lungimea de und\u0103). Pentru a ob\u021bine o imagine RGB, a fost necesar, de asemenea, s\u0103 se creeze o re\u021bea de sticl\u0103 colorat\u0103 pe \u00eentreaga matrice. Cea mai popular\u0103 implementare a sa se nume\u0219te filtru Bayer \u0219i este utilizat\u0103 ast\u0103zi \u00een majoritatea senzorilor. Arat\u0103 ca \u00een imaginea de mai jos.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/7dc43fed6c8080917d09275c23cd614c.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Astfel, fiecare pixel al matricei capteaz\u0103 doar componenta R, G sau B, iar ceilal\u021bi fotoni sunt pur \u0219i simplu reflecta\u021bi de filtrul Bayer. Componenta lips\u0103 este estimat\u0103 printr-o medie simpl\u0103 a valorilor pixelilor vecini.<\/p>\n<p>Celulele verzi din filtrul Bayer sunt mai numeroase, acesta fiind inspirat de ochiul uman. Astfel, din cele 50 de milioane de pixeli de pe senzor, culoarea verde va fi captat\u0103 de 25 milioane, iar ro\u0219u \u0219i albastru de c\u00e2te 12,5 milioane. Restul va fi mediatizat; acest proces se nume\u0219te debayerizare sau demosaicare, \u0219i acesta este un mecanism amuzant pe care se bazeaz\u0103 totul.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/aa2d9d74153b26e3f0cdb6d2ba954fc2.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p style=\"text-align: center;\">De fapt, fiecare matrice are propriul algoritm de demosaicing patentat, dar \u00een cadrul acestei pove\u0219ti vom trece cu vederea acest aspect.<\/p>\n<p style=\"text-align: center;\">\n<p>Alte tipuri de matrice (de tip Foveon) nu s-au impus deocamdat\u0103. Totu\u0219i, unele m\u0103rci \u00eencearc\u0103 s\u0103 utilizeze matrice f\u0103r\u0103 filtrul Bayer pentru a \u00eembun\u0103t\u0103\u021bi claritatea \u0219i gama dinamic\u0103.<\/p>\n<p>C\u00e2nd lumina este slab\u0103 sau detaliile obiectului sunt foarte mici, pierdem o mul\u021bime de informa\u021bii, deoarece filtrul Bayer taie cu neru\u0219inare fotonii cu lungimi de und\u0103 nedorite. De aceea a fost inventat\u0103 tehnologia Pixel Shifting \u2013 mutarea matricei cu 1 pixel sus, jos, dreapta sau st\u00e2nga, pentru a le captura pe toate. Fotografia nu devine de 4 ori mai mare, a\u0219a cum ar p\u0103rea, ci procesorul folose\u0219te aceste date pentru a \u00eenregistra mai precis valoarea fiec\u0103rui pixel. Mediaz\u0103 nu pe baza vecinilor, ci pe cele patru valori ale sale.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/f0c5f4eca43894ef480d1573ef4fddda.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Tremurul m\u00e2inilor noastre atunci c\u00e2nd facem fotografii cu telefonul face ca acest proces s\u0103 fie o consecin\u021b\u0103 natural\u0103. \u00cen ultimele versiuni ale Google Pixel, aceast\u0103 func\u021bionalitate este implementat\u0103 \u0219i se activeaz\u0103 \u00eentotdeauna c\u00e2nd folose\u0219ti zoom-ul telefonului, \u2013 se nume\u0219te Super Res Zoom (da, \u00eemi place \u0219i mie denumirea lor necru\u021b\u0103toare). Chinezii au copiat-o \u0219i \u00een telefoanele lor, de\u0219i a ie\u0219it pu\u021bin mai prost.<\/p>\n<p>Suprapunerea unor fotografii u\u0219or deplasate permite colectarea mai multor informa\u021bii despre culoarea fiec\u0103rui pixel, astfel \u00eenc\u00e2t se reduce zgomotul, se cre\u0219te claritatea \u0219i se m\u0103re\u0219te rezolu\u021bia f\u0103r\u0103 a cre\u0219te num\u0103rul fizic de megapixeli al matricei. Majoritatea flagship-urilor Android moderne fac acest lucru automat, f\u0103r\u0103 ca utilizatorii lor s\u0103 se g\u00e2ndeasc\u0103 la asta.<\/p>\n<h2><noindex><a rel=\"nofollow\" href=\"#contents\" class=\"right paragraph tooltip\" data-tooltip=\"\u0412\u0435\u0440\u043d\u0443\u0442\u044c\u0441\u044f \u043a \u043e\u0433\u043b\u0430\u0432\u043b\u0435\u043d\u0438\u044e\">\ud83d\udd3d<\/a><\/noindex><noindex><a rel=\"nofollow\" name=\"\u0421\u0442\u0435\u043a\u0438\u043d\u0433 \u043f\u043e \u0444\u043e\u043a\u0443\u0441\u0443 &mdash; \u043b\u044e\u0431\u0430\u044f \u0433\u043b\u0443\u0431\u0438\u043d\u0430 \u0440\u0435\u0437\u043a\u043e\u0441\u0442\u0438 \u0438 \u0440\u0435\u0444\u043e\u043a\u0443\u0441 \u0432 \u043f\u043e\u0441\u0442\u043f\u0440\u043e\u0434\u0430\u043a\u0448\u0435\u043d\u0435\" href=\"#\u0421\u0442\u0435\u043a\u0438\u043d\u0433 \u043f\u043e \u0444\u043e\u043a\u0443\u0441\u0443 &mdash; \u043b\u044e\u0431\u0430\u044f \u0433\u043b\u0443\u0431\u0438\u043d\u0430 \u0440\u0435\u0437\u043a\u043e\u0441\u0442\u0438 \u0438 \u0440\u0435\u0444\u043e\u043a\u0443\u0441 \u0432 \u043f\u043e\u0441\u0442\u043f\u0440\u043e\u0434\u0430\u043a\u0448\u0435\u043d\u0435\" class=\"right paragraph tooltip\" data-tooltip=\"\u0421\u0441\u044b\u043b\u043a\u0430 \u043d\u0430 \u044d\u0442\u0443 \u0433\u043b\u0430\u0432\u0443\">#<\/a><\/noindex>Stakingul pe focalizare \u2013 orice ad\u00e2ncime de c\u00e2mp \u0219i refocusare \u00een post-produc\u021bie.<\/h2>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/19299836bf612a9425c5633572d7879c.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Metoda provine din macrofotografie, unde o ad\u00e2ncime mic\u0103 a c\u00e2mpului de focalizare a fost \u00eentotdeauna o problem\u0103. Pentru a avea \u00eentregul obiect \u00een focalizare, era necesar s\u0103 se fac\u0103 mai multe fotografii cu schimbarea focaliz\u0103rii \u00eenainte-\u00eenapoi, pentru a le lipi apoi \u00eentr-o imagine clar\u0103. Aceast\u0103 metod\u0103 a fost folosit\u0103 frecvent \u0219i de amatorii fotografiei de peisaj, f\u0103c\u00e2nd clar at\u00e2t planul din fa\u021b\u0103, c\u00e2t \u0219i pe cel din spate.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/a2336f7ce07d4bbd931881b2b8a16387.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Toate acestea s-au mutat \u0219i pe smartphone-uri, de\u0219i f\u0103r\u0103 prea mult hype. \u00cen 2013, apare Nokia Lumia 1020 cu \u201eRefocus App\u201d, iar \u00een 2014, Samsung Galaxy S5 cu modul \u201eSelective Focus\u201d. Funcionau pe aceea\u0219i schem\u0103: prin ap\u0103sarea unui buton, f\u0103ceau rapid 3 fotografii \u2013 una cu focus \u201enormal\u201d, a doua cu focus mutat \u00een fa\u021b\u0103 \u0219i a treia cu focus mutat \u00een spate. Programul aliniaz\u0103 cadrele \u0219i permite alegerea uneia dintre ele, ceea ce era prezentat ca o \u201eadev\u0103rat\u0103\u201d controlare a focaliz\u0103rii \u00een post-produc\u021bie.<\/p>\n<p>Nu a fost necesar\u0103 nicio procesare suplimentar\u0103, deoarece chiar \u0219i acest simplu hack a fost suficient pentru a mai \u00eenfinge un cui \u00een capacul Lytro \u0219i al analogilor s\u0103i cu refocusare real\u0103. Apropo, s\u0103 discut\u0103m despre ei (masterii tranzi\u021biilor de nivel 80).<\/p>\n<h2><noindex><a rel=\"nofollow\" href=\"#contents\" class=\"right paragraph tooltip\" data-tooltip=\"\u0412\u0435\u0440\u043d\u0443\u0442\u044c\u0441\u044f \u043a \u043e\u0433\u043b\u0430\u0432\u043b\u0435\u043d\u0438\u044e\">\ud83d\udd3d<\/a><\/noindex><noindex><a rel=\"nofollow\" name=\"\u0412\u044b\u0447\u0438\u0441\u043b\u0438\u0442\u0435\u043b\u044c\u043d\u044b\u0435 \u043c\u0430\u0442\u0440\u0438\u0446\u044b &mdash; \u0441\u0432\u0435\u0442\u043e\u0432\u044b\u0435 \u043f\u043e\u043b\u044f \u0438 \u043f\u043b\u0435\u043d\u043e\u043f\u0442\u0438\u043a\u0430\" href=\"#\u0412\u044b\u0447\u0438\u0441\u043b\u0438\u0442\u0435\u043b\u044c\u043d\u044b\u0435 \u043c\u0430\u0442\u0440\u0438\u0446\u044b &mdash; \u0441\u0432\u0435\u0442\u043e\u0432\u044b\u0435 \u043f\u043e\u043b\u044f \u0438 \u043f\u043b\u0435\u043d\u043e\u043f\u0442\u0438\u043a\u0430\" class=\"right paragraph tooltip\" data-tooltip=\"\u0421\u0441\u044b\u043b\u043a\u0430 \u043d\u0430 \u044d\u0442\u0443 \u0433\u043b\u0430\u0432\u0443\">#<\/a><\/noindex>Matricele computa\u021bionale \u2013 c\u00e2mpuri de lumin\u0103 \u0219i plenoptic\u0103.<\/h2>\n<p>Cum am \u00een\u021beles anterior, matricele noastre sunt un co\u0219mar pe coclogi. Ne-am obi\u0219nuit \u0219i \u00eencerc\u0103m s\u0103 tr\u0103im cu asta. Din punct de vedere structural, ele nu s-au schimbat mult de la \u00eenceputul vremurilor. Am perfec\u021bionat doar procesul tehnologic \u2013 am mic\u0219orat distan\u021ba dintre pixeli, am luptat cu zgomotele de fond, am ad\u0103ugat pixeli speciali pentru a func\u021biona autofocusul de faz\u0103. Dar ia chiar \u0219i cea mai scump\u0103 camer\u0103 reflex \u0219i \u00eencearc\u0103 s\u0103 fotografiezi un pisoi alerg\u0103tor \u00een lumin\u0103 de camer\u0103 \u2013 pisoiul, cu bl\u00e2nde\u021be, va c\u00e2\u0219tiga.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/dcb87026c62ed4604bc67a4455fa0997.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>De mult timp \u00eencerc\u0103m s\u0103 invent\u0103m ceva mai bun. O mul\u021bime de \u00eencerc\u0103ri \u0219i cercet\u0103ri \u00een acest domeniu pot fi g\u0103site c\u0103ut\u00e2nd \u201esensor computa\u021bional\u201d sau \u201esensor non-Bayer\u201d, iar chiar exemplul cu Pixel Shifting de mai sus poate fi considerat o \u00eencercare de \u00eembun\u0103t\u0103\u021bire a senzorilor prin metode computa\u021bionale. Totu\u0219i, cele mai promi\u021b\u0103toare pove\u0219ti din ultimele dou\u0103zeci de ani provin tocmai din lumea a\u0219a-numitelor camere plenoptice.<\/p>\n<p>Ca s\u0103 nu adormiti de anticipa\u021bia cuvintelor complicate care urmeaz\u0103, v\u0103 \u00eemp\u0103rt\u0103\u0219esc un insight: camera celor mai recente Google Pixel este oarecum \u201epleno-optic\u0103\u201d. Cu doar dou\u0103 pixeli \u00een plus, dar chiar \u0219i at\u00e2t \u00eei permite s\u0103 calculeze o ad\u00e2ncime optic\u0103 corect\u0103 a cadrului, f\u0103r\u0103 a avea nevoie de o a doua camer\u0103, a\u0219a cum au toate celelalte.<\/p>\n<p>Plenoptica \u2014 o arm\u0103 puternic\u0103, care \u00eenc\u0103 nu a fost utilizat\u0103. Voi oferi un link c\u0103tre unuia dintre cele mai recente articole preferate. <noindex><a rel=\"nofollow noopener noreferrer\" href=\"https:\/\/m.habr.com\/ru\/post\/440652\/\" target=\"_blank\">despre capacit\u0103\u021bile camerelor plenoptice \u0219i viitorul nostru cu acestea<\/a><\/noindex>, de unde am preluat exemplele.<\/p>\n<p class=\"h3\"><noindex><a rel=\"nofollow\" href=\"#contents\" class=\"right paragraph tooltip\" data-tooltip=\"\u0412\u0435\u0440\u043d\u0443\u0442\u044c\u0441\u044f \u043a \u043e\u0433\u043b\u0430\u0432\u043b\u0435\u043d\u0438\u044e\">\ud83d\udd3d<\/a><\/noindex><noindex><a rel=\"nofollow\" name=\"\u041f\u043b\u0435\u043d\u043e\u043f\u0442\u0438\u0447\u0435\u0441\u043a\u0430\u044f \u043a\u0430\u043c\u0435\u0440\u0430 &mdash; \u0441\u043a\u043e\u0440\u043e \u0431\u0443\u0434\u0435\u0442 \u043a\u0430\u0436\u0434\u0430\u044f\" href=\"#\u041f\u043b\u0435\u043d\u043e\u043f\u0442\u0438\u0447\u0435\u0441\u043a\u0430\u044f \u043a\u0430\u043c\u0435\u0440\u0430 &mdash; \u0441\u043a\u043e\u0440\u043e \u0431\u0443\u0434\u0435\u0442 \u043a\u0430\u0436\u0434\u0430\u044f\" class=\"right paragraph tooltip\" data-tooltip=\"\u0421\u0441\u044b\u043b\u043a\u0430 \u043d\u0430 \u044d\u0442\u0443 \u0433\u043b\u0430\u0432\u0443\">#<\/a><\/noindex><\/p>\n<h2>Camera plenoptic\u0103 \u2014 cur\u00e2nd va fi disponibil\u0103 pentru toat\u0103 lumea.<\/h2>\n<p>Inventat\u0103 \u00een 1994, asamblat\u0103 la Stanford \u00een 2004. Prima camer\u0103 de consum \u2014 Lytro, lansat\u0103 \u00een 2012. Industria VR experimenteaz\u0103 activ acum cu tehnologii similare.<\/p>\n<p>Diferen\u021ba dintre o camer\u0103 obi\u0219nuit\u0103 \u0219i una plenoptic\u0103 const\u0103 \u00een faptul c\u0103 senzorul este acoperit cu o re\u021bea de lentile, fiecare dintre acestea acoperind c\u00e2\u021biva pixeli reali. A\u0219a arat\u0103:<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/209d4729485d9f36976c47d7b93ee6f7.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Dac\u0103 se calculeaz\u0103 corect distan\u021ba de la re\u021bea la senzor \u0219i dimensiunea diafragmei, imaginea final\u0103 va avea clustere clare de pixeli \u2014 mici versiuni ale imaginii originale.<\/p>\n<p>Se pare c\u0103 dac\u0103 lu\u0103m din fiecare cluster, s\u0103 zicem, un pixel central \u0219i unim imaginea doar pe baza acestora \u2014 nu va diferi de cea realizat\u0103 cu o camer\u0103 obi\u0219nuit\u0103. Da, am pierdut pu\u021bin la rezolu\u021bie, dar putem s\u0103-i cerem lui Sony s\u0103 adauge c\u00e2\u021biva megapixeli \u00een noile senzori.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/290518da93056e53314ed8219709f9a2.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Distrac\u021bia abia acum \u00eencepe. Dac\u0103 lu\u0103m un alt pixel din fiecare cluster \u0219i unim din nou imaginea \u2014 vom ob\u021bine din nou o fotografie normal\u0103, doar c\u0103, parca, realizat\u0103 cu un pixel shiftat. Astfel, av\u00e2nd clustere de 10 \u00d7 10 pixeli, vom ob\u021bine 100 de imagini ale obiectului din \u201epu\u021bin\u201d unghiuri diferite.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/70bbfa990d697dfb5b2c119853a8810e.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Cu c\u00e2t dimensiunea clusterului este mai mare \u2014 cu at\u00e2t mai multe imagini, dar cu rezolu\u021bie mai mic\u0103. \u00cen lumea smartphone-urilor cu senzori de 41 megapixeli, putem s\u0103 ne permitem s\u0103 ignor\u0103m pu\u021bin rezolu\u021bia, dar totul are o limit\u0103. Trebuie s\u0103 men\u021binem un echilibru.<\/p>\n<p>Bine, am asamblat o camer\u0103 plenoptic\u0103, \u0219i ce ne ofer\u0103 aceasta?<\/p>\n<p><strong>Refocalizare cinstit\u0103<\/strong><\/p>\n<p>Func\u021bia despre care au vorbit to\u021bi jurnali\u0219tii \u00een articolele despre Lytro este capacitatea de a corecta focalizarea \u00een post-produc\u021bie. Prin corectare cinstit\u0103 se \u00een\u021belege c\u0103 nu aplic\u0103m diverse algoritmi de deblurare, ci folosim exclusiv pixelii disponibili, aleg\u00e2ndu-i sau mediatiz\u00e2ndu-i din clustere \u00een ordinea necesar\u0103.<\/p>\n<p>O fotografie RAW de la o camer\u0103 plenoptic arat\u0103 ciudat. Pentru a ob\u021bine un JPEG clar obi\u0219nuit, trebuie mai \u00eent\u00e2i s\u0103 o reconstituim. Asta \u00eenseamn\u0103 c\u0103 trebuie s\u0103 alegem fiecare pixel din JPEG din unul dintre clusterele RAW. \u00cen func\u021bie de modul \u00een care \u00eei alegem, rezultatul va varia.<\/p>\n<p>De exemplu, cu c\u00e2t un cluster este mai departe de locul unde a c\u0103zut raz\u0103 original\u0103, cu at\u00e2t aceast\u0103 raz\u0103 devine mai neclar\u0103. Asta din cauza opticii. Pentru a ob\u021bine o imagine cu o focalizare difuz\u0103, trebuie doar s\u0103 alegem pixelii la distan\u021ba dorit\u0103 de original - fie mai aproape, fie mai departe.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/2c091775113b80bc0333624190529ffe.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>&nbsp;<\/p>\n<p>A fost mai complicat s\u0103 ajust\u0103m focalizarea asupra noastr\u0103 - din punct de vedere fizic, erau mai pu\u021bini pixeli \u00een clustere. La \u00eenceput, dezvoltatorii nici nu au vrut s\u0103 ofere utilizatorului posibilitatea de a focaliza manual - camera decidea aceasta programatic. Utilizatorilor nu le-a pl\u0103cut acest viitor, a\u0219a c\u0103 func\u021bia a fost ad\u0103ugat\u0103 \u00een versiuni ulterioare ale firmware-ului sub numele de \u201emod creativ\u201d, darau f\u0103cut refocalizarea foarte limitat\u0103 exact din acest motiv.<\/p>\n<p><strong>Harta de profunzime \u0219i 3D cu o singur\u0103 camer\u0103&nbsp;&nbsp;&nbsp;<\/strong><\/p>\n<p>Una dintre cele mai simple opera\u021biuni \u00een plenoptic este ob\u021binerea unei h\u0103r\u021bi de ad\u00e2ncime. Pentru aceasta, trebuie doar s\u0103 reconstituim dou\u0103 cadre diferite \u0219i s\u0103 calcul\u0103m c\u00e2t de mult sunt deplasate obiectele \u00een ele. Cu c\u00e2t deplasarea este mai mare, cu at\u00e2t obiectul este mai departe de camer\u0103.<\/p>\n<p>Recent, Google a cump\u0103rat \u0219i a \u00eenchis Lytro, dar a folosit tehnologiile lor pentru VR-ul s\u0103u \u0219i... pentru camera din Pixel. \u00cencep\u00e2nd cu Pixel 2, camera a devenit pentru prima dat\u0103 \u201epu\u021bin\u201d plenoptic, de fapt, cu clustere ce con\u021bin doar doi pixeli. Aceasta a permis Google-ului s\u0103 nu instaleze o a doua camer\u0103, cum au f\u0103cut ceilal\u021bi, ci s\u0103 calculeze harta de ad\u00e2ncime exclusiv pe baza unei singure fotografii.<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/11212c1d7ad0ad9bca448c01b81e49fc.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/4112733d57496d546c66100c541dee22.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Harta ad\u00e2ncimii se construie\u0219te pe baza a dou\u0103 cadre, deplasate cu un subpixeli. Acest lucru este suficient pentru a calcula harta binar\u0103 a ad\u00e2ncimii \u0219i a separa primul plan de fundal, estomp\u00e2nd ultimul \u00een mod popular \u00een bokeh. Rezultatul acestei stratific\u0103ri este apoi netezit \u0219i \u201e\u00eembun\u0103t\u0103\u021bit\u201d de re\u021bele neuronale, care sunt antrenate s\u0103 \u00eembun\u0103t\u0103\u021beasc\u0103 h\u0103r\u021bile de ad\u00e2ncime (\u0219i nu s\u0103 estompeze, a\u0219a cum mul\u021bi cred).<\/p>\n<p><img decoding=\"async\" alt=\"Articol nou: Fotografia Computa\u021bional\u0103\" src=\"\/wp-content\/uploads\/2019\/06\/f94ae3d368c97cf0e5c63c288dfaab23.jpg\" style=\"display:block;margin: 0 auto;\" \/><\/p>\n<p>Un alt aspect este c\u0103 plenoptica a ajuns la smartphone-uri aproape f\u0103r\u0103 costuri. Deja foloseam lentile pe aceste matri\u021be mici pentru a cre\u0219te fluxul de lumin\u0103. \u00cen urm\u0103toarele Pixel, Google pl\u0103nuie\u0219te s\u0103 mearg\u0103 mai departe \u0219i s\u0103 acopere patru fotodiodi cu o lentil\u0103.<\/p>\n<p>Sursa: <a content=\"nofollow\" rel=\"nofollow\" href=\"https:\/\/3dnews.ru\/989337\">3dnews.ru<\/a><\/p>","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"excerpt":{"rendered":"<p>\u041e\u0440\u0438\u0433\u0438\u043d\u0430\u043b \u0441\u0442\u0430\u0442\u044c\u0438 \u0440\u0430\u0437\u043c\u0435\u0449\u0435\u043d \u043d\u0430 \u0441\u0430\u0439\u0442\u0435 \u0412\u0430\u0441\u0442\u0440\u0438\u043a.\u0440\u0443&nbsp;\u0438 \u043e\u043f\u0443\u0431\u043b\u0438\u043a\u043e\u0432\u0430\u043d \u043d\u0430 3DNews \u0441 \u0440\u0430\u0437\u0440\u0435\u0448\u0435\u043d\u0438\u044f \u0430\u0432\u0442\u043e\u0440\u0430. \u041c\u044b \u043f\u0440\u0438\u0432\u043e\u0434\u0438\u043c \u043f\u043e\u043b\u043d\u044b\u0439 \u0442\u0435\u043a\u0441\u0442 \u0441\u0442\u0430\u0442\u044c\u0438, \u0437\u0430 \u0438\u0441\u043a\u043b\u044e\u0447\u0435\u043d\u0438\u0435\u043c \u043e\u0433\u0440\u043e\u043c\u043d\u043e\u0433\u043e \u043a\u043e\u043b\u0438\u0447\u0435\u0441\u0442\u0432\u0430 \u0441\u0441\u044b\u043b\u043e\u043a &mdash; \u043e\u043d\u0438 \u043f\u0440\u0438\u0433\u043e\u0434\u044f\u0442\u0441\u044f \u0442\u0435\u043c, \u043a\u0442\u043e \u0432\u0441\u0435\u0440\u044c\u0435\u0437 \u0437\u0430\u0438\u043d\u0442\u0435\u0440\u0435\u0441\u043e\u0432\u0430\u043b\u0441\u044f \u0442\u0435\u043c\u043e\u0439 \u0438 \u0445\u043e\u0442\u0435\u043b \u0431\u044b \u0438\u0437\u0443\u0447\u0438\u0442\u044c \u0442\u0435\u043e\u0440\u0435\u0442\u0438\u0447\u0435\u0441\u043a\u0438\u0435 \u0430\u0441\u043f\u0435\u043a\u0442\u044b \u0432\u044b\u0447\u0438\u0441\u043b\u0438\u0442\u0435\u043b\u044c\u043d\u043e\u0439 \u0444\u043e\u0442\u043e\u0433\u0440\u0430\u0444\u0438\u0438 \u0431\u043e\u043b\u0435\u0435 \u0433\u043b\u0443\u0431\u043e\u043a\u043e, \u043d\u043e \u0434\u043b\u044f \u0448\u0438\u0440\u043e\u043a\u043e\u0439 \u0430\u0443\u0434\u0438\u0442\u043e\u0440\u0438\u0438 \u043c\u044b \u0441\u043e\u0447\u043b\u0438 \u044d\u0442\u043e\u0442 \u043c\u0430\u0442\u0435\u0440\u0438\u0430\u043b \u0438\u0437\u0431\u044b\u0442\u043e\u0447\u043d\u044b\u043c.&nbsp;&nbsp; \u0421\u0435\u0433\u043e\u0434\u043d\u044f \u043d\u0438 \u043e\u0434\u043d\u0430 \u043f\u0440\u0435\u0437\u0435\u043d\u0442\u0430\u0446\u0438\u044f \u0441\u043c\u0430\u0440\u0442\u0444\u043e\u043d\u0430 \u043d\u0435 [&hellip;]<\/p>\n","protected":false,"gt_translate_keys":[{"key":"rendered","format":"html"}]},"author":1,"featured_media":26610,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[702],"tags":[],"class_list":["post-35473","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-news"],"aioseo_notices":[],"aioseo_head":"\n\t\t<!-- All in One SEO 5.0.2.1 - aioseo.com -->\n\t<meta name=\"description\" content=\"\u041e\u0440\u0438\u0433\u0438\u043d\u0430\u043b \u0441\u0442\u0430\u0442\u044c\u0438 \u0440\u0430\u0437\u043c\u0435\u0449\u0435\u043d \u043d\u0430 \u0441\u0430\u0439\u0442\u0435 \u0412\u0430\u0441\u0442\u0440\u0438\u043a.\u0440\u0443 \u0438 \u043e\u043f\u0443\u0431\u043b\u0438\u043a\u043e\u0432\u0430\u043d \u043d\u0430 3DNews \u0441 \u0440\u0430\u0437\u0440\u0435\u0448\u0435\u043d\u0438\u044f \u0430\u0432\u0442\u043e\u0440\u0430.\" \/>\n\t<meta name=\"robots\" content=\"max-image-preview:large\" \/>\n\t<meta name=\"author\" content=\"Yuri Gagarin\"\/>\n\t<link rel=\"canonical\" href=\"https:\/\/prohoster.info\/ro\/blog\/news\/novaya-statya-vychislitelnaya-fotografiya\" \/>\n\t<meta name=\"generator\" content=\"All in One SEO (AIOSEO) 5.0.2.1\" \/>\n\t\t<meta property=\"og:locale\" content=\"ro_RO\" \/>\n\t\t<meta property=\"og:site_name\" content=\"ProHoster | \u041a\u0443\u043f\u0438\u0442\u044c \u043d\u0430\u0434\u0435\u0436\u043d\u044b\u0439 \u0445\u043e\u0441\u0442\u0438\u043d\u0433 \u0434\u043b\u044f \u0441\u0430\u0439\u0442\u043e\u0432 \u0441 \u0437\u0430\u0449\u0438\u0442\u043e\u0439 \u043e\u0442 DDoS, VPS VDS \u0441\u0435\u0440\u0432\u0435\u0440\u044b\" \/>\n\t\t<meta property=\"og:type\" content=\"article\" \/>\n\t\t<meta property=\"og:title\" content=\"\ud83e\udd47\u041d\u043e\u0432\u0430\u044f \u0441\u0442\u0430\u0442\u044c\u044f: \u0412\u044b\u0447\u0438\u0441\u043b\u0438\u0442\u0435\u043b\u044c\u043d\u0430\u044f \u0424\u043e\u0442\u043e\u0433\u0440\u0430\u0444\u0438\u044f | ProHoster\" \/>\n\t\t<meta property=\"og:description\" content=\"\u041e\u0440\u0438\u0433\u0438\u043d\u0430\u043b \u0441\u0442\u0430\u0442\u044c\u0438 \u0440\u0430\u0437\u043c\u0435\u0449\u0435\u043d \u043d\u0430 \u0441\u0430\u0439\u0442\u0435 \u0412\u0430\u0441\u0442\u0440\u0438\u043a.\u0440\u0443 \u0438 \u043e\u043f\u0443\u0431\u043b\u0438\u043a\u043e\u0432\u0430\u043d \u043d\u0430 3DNews \u0441 \u0440\u0430\u0437\u0440\u0435\u0448\u0435\u043d\u0438\u044f \u0430\u0432\u0442\u043e\u0440\u0430.\" \/>\n\t\t<meta property=\"og:url\" content=\"https:\/\/prohoster.info\/ro\/blog\/news\/novaya-statya-vychislitelnaya-fotografiya\" \/>\n\t\t<meta property=\"og:image\" content=\"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg\" \/>\n\t\t<meta property=\"og:image:secure_url\" content=\"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg\" \/>\n\t\t<meta property=\"og:image:width\" content=\"350\" \/>\n\t\t<meta property=\"og:image:height\" content=\"350\" \/>\n\t\t<meta property=\"article:published_time\" content=\"2019-10-31T19:04:32+00:00\" \/>\n\t\t<meta property=\"article:modified_time\" content=\"2019-10-31T19:04:32+00:00\" \/>\n\t\t<meta property=\"article:publisher\" content=\"https:\/\/www.facebook.com\/prohoster\" \/>\n\t\t<meta property=\"article:author\" content=\"https:\/\/www.facebook.com\/prohoster\" \/>\n\t\t<!-- All in One SEO -->\n\n","aioseo_head_json":{"title":"\ud83e\udd47Articol nou: Fotografia Computa\u021bional\u0103 | ProHoster","description":"Articolul original este publicat pe site-ul Vastrik.ru \u0219i publicat pe 3DNews cu permisiunea autorului.","canonical_url":"https:\/\/prohoster.info\/ro\/blog\/news\/novaya-statya-vychislitelnaya-fotografiya","robots":"max-image-preview:large","keywords":"","webmasterTools":{"miscellaneous":""},"schema":null,"og:locale":"ro_RO","og:site_name":"ProHoster | \u041a\u0443\u043f\u0438\u0442\u044c \u043d\u0430\u0434\u0435\u0436\u043d\u044b\u0439 \u0445\u043e\u0441\u0442\u0438\u043d\u0433 \u0434\u043b\u044f \u0441\u0430\u0439\u0442\u043e\u0432 \u0441 \u0437\u0430\u0449\u0438\u0442\u043e\u0439 \u043e\u0442 DDoS, VPS VDS \u0441\u0435\u0440\u0432\u0435\u0440\u044b","og:type":"article","og:title":"\ud83e\udd47\u041d\u043e\u0432\u0430\u044f \u0441\u0442\u0430\u0442\u044c\u044f: \u0412\u044b\u0447\u0438\u0441\u043b\u0438\u0442\u0435\u043b\u044c\u043d\u0430\u044f \u0424\u043e\u0442\u043e\u0433\u0440\u0430\u0444\u0438\u044f | ProHoster","og:description":"\u041e\u0440\u0438\u0433\u0438\u043d\u0430\u043b \u0441\u0442\u0430\u0442\u044c\u0438 \u0440\u0430\u0437\u043c\u0435\u0449\u0435\u043d \u043d\u0430 \u0441\u0430\u0439\u0442\u0435 \u0412\u0430\u0441\u0442\u0440\u0438\u043a.\u0440\u0443 \u0438 \u043e\u043f\u0443\u0431\u043b\u0438\u043a\u043e\u0432\u0430\u043d \u043d\u0430 3DNews \u0441 \u0440\u0430\u0437\u0440\u0435\u0448\u0435\u043d\u0438\u044f \u0430\u0432\u0442\u043e\u0440\u0430.","og:url":"https:\/\/prohoster.info\/ro\/blog\/news\/novaya-statya-vychislitelnaya-fotografiya","og:image":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:secure_url":"https:\/\/prohoster.info\/wp-content\/uploads\/2021\/11\/logo-350.jpg","og:image:width":350,"og:image:height":350,"article:published_time":"2019-10-31T19:04:32+00:00","article:modified_time":"2019-10-31T19:04:32+00:00","article:publisher":"https:\/\/www.facebook.com\/prohoster","article:author":"https:\/\/www.facebook.com\/prohoster"},"aioseo_meta_data":{"post_id":"35473","title":null,"description":null,"keywords":null,"keyphrases":null,"primary_term":null,"canonical_url":null,"og_title":null,"og_description":null,"og_object_type":"default","og_image_type":"default","og_image_url":null,"og_image_width":null,"og_image_height":null,"og_image_custom_url":null,"og_image_custom_fields":null,"og_video":null,"og_custom_url":null,"og_article_section":null,"og_article_tags":null,"twitter_use_og":false,"twitter_card":"default","twitter_image_type":"default","twitter_image_url":null,"twitter_image_custom_url":null,"twitter_image_custom_fields":null,"twitter_title":null,"twitter_description":null,"schema":{"blockGraphs":[],"customGraphs":[],"default":{"data":{"Article":[],"Course":[],"Dataset":[],"FAQPage":[],"Movie":[],"Person":[],"Product":[],"ProductReview":[],"Car":[],"Recipe":[],"Service":[],"SoftwareApplication":[],"WebPage":[]},"graphName":"","isEnabled":true},"graphs":[]},"schema_type":null,"schema_type_options":null,"pillar_content":false,"robots_default":true,"robots_noindex":false,"robots_noarchive":false,"robots_nosnippet":false,"robots_nofollow":false,"robots_noimageindex":false,"robots_noodp":false,"robots_notranslate":false,"robots_max_snippet":null,"robots_max_videopreview":null,"robots_max_imagepreview":"large","priority":null,"frequency":null,"local_seo":null,"seo_analyzer_scan_date":"2026-01-21 23:25:19","breadcrumb_settings":null,"limit_modified_date":false,"reviewed_by":null,"ai":null,"created":"2021-03-01 02:01:27","updated":"2026-01-21 23:25:19","focus_keyword":null,"additional_keywords":null,"truseo_locale":null},"gt_translate_keys":[{"key":"link","format":"url"}],"_links":{"self":[{"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/posts\/35473","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/comments?post=35473"}],"version-history":[{"count":0,"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/posts\/35473\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/media\/26610"}],"wp:attachment":[{"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/media?parent=35473"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/categories?post=35473"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/prohoster.info\/ro\/wp-json\/wp\/v2\/tags?post=35473"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}