{"id":959,"date":"2023-03-14T15:55:04","date_gmt":"2023-03-14T12:55:04","guid":{"rendered":"https:\/\/legendhomework.com\/tr\/?p=959"},"modified":"2023-03-14T15:55:04","modified_gmt":"2023-03-14T12:55:04","slug":"goruntuyu-test-etme-multimedya-bolumu-multimedya-bolumu-odevleri-multimedya-bolumu-tez-yaptirma-multimedya-bolumu-odev-ucretleri","status":"publish","type":"post","link":"https:\/\/legendhomework.com\/tr\/goruntuyu-test-etme-multimedya-bolumu-multimedya-bolumu-odevleri-multimedya-bolumu-tez-yaptirma-multimedya-bolumu-odev-ucretleri\/","title":{"rendered":"G\u00f6r\u00fcnt\u00fcy\u00fc Test Etme \u2013 Multimedya B\u00f6l\u00fcm\u00fc \u2013 Multimedya B\u00f6l\u00fcm\u00fc \u00d6devleri \u2013 Multimedya B\u00f6l\u00fcm\u00fc Tez Yapt\u0131rma \u2013Multimedya B\u00f6l\u00fcm\u00fc \u00d6dev \u00dccretleri"},"content":{"rendered":"

G\u00f6r\u00fcnt\u00fcy\u00fc Test Etme<\/span><\/strong><\/h3>\n

\u0130lk \u00f6zellik, bir y\u00fcz g\u00f6r\u00fcnt\u00fcs\u00fc dikkate al\u0131narak \u00f6rneklenebilir. Ki\u015finin yanaklar\u0131ndaki ve aln\u0131ndaki yerel kom\u015fuluklar, ba\u011f\u0131ms\u0131z olarak bak\u0131ld\u0131\u011f\u0131nda genellikle ay\u0131rt edilemez. Bu kar\u015f\u0131la\u015ft\u0131rmada ayd\u0131nlatma ve di\u011fer \u201c\u00f6zel\u201d niteliklerin (yara izleri, benler, do\u011fum lekeleri vb.) etkilerinin bulunmad\u0131\u011f\u0131n\u0131 varsayd\u0131k.<\/span><\/p>\n

Bu g\u00f6zlemi test etmenin kolay bir y\u00f6ntemi, bir g\u00f6r\u00fcnt\u00fcdeki bu benzer g\u00f6r\u00fcnt\u00fc k\u0131s\u0131mlar\u0131n\u0131 bulmak ve yeni bir g\u00f6r\u00fcnt\u00fc olu\u015fturmak i\u00e7in bunlar\u0131 rasgele de\u011fi\u015ftirmektir. Yeni g\u00f6r\u00fcnt\u00fc orijinaline benziyorsa, g\u00f6zlemimiz do\u011frudur.<\/span><\/p>\n

Benzer \u015fekilde, belirli bir \u00f6zellik sergileyen t\u00fcm mahalleler pratik olarak ayn\u0131 \u015fekilde ele al\u0131nabilir. Bu mahalleler, ayn\u0131 istatistiksel s\u00fcre\u00e7 taraf\u0131ndan \u00fcretilmi\u015f olarak kabul edilebilir. Mahalleler i\u00e7in hedeflerin, kullan\u0131lan her model i\u00e7in bir istatistiksel s\u00fcrecin ortalamas\u0131 olarak yorumlanabilece\u011fi g\u00f6sterilmi\u015ftir.<\/span><\/p>\n

Bu ilk \u00f6zelli\u011fin \u00f6rnekleri resimlerde bol miktarda bulunur. Son zamanlarda s\u0131k\u0131\u015ft\u0131rma kazan\u00e7lar\u0131n\u0131 art\u0131rmak i\u00e7in kullan\u0131ld\u0131. Kay\u0131pl\u0131 g\u00f6r\u00fcnt\u00fc s\u0131k\u0131\u015ft\u0131rma i\u00e7in standart \u015femalar, genellikle daha verimli ve derli toplu bir \u015fekilde tan\u0131mlanabilecek k\u00fc\u00e7\u00fck g\u00f6r\u00fcnt\u00fc bloklar\u0131nda bulunan y\u00fcksek derecede fazlal\u0131kl\u0131 bilgilere dayan\u0131r.<\/span><\/p>\n

Temel bile\u015fen analizi (PCA) yoluyla s\u0131k\u0131\u015ft\u0131rma durumunda, g\u00f6r\u00fcnt\u00fcdeki t\u00fcm bloklar i\u00e7in tek bir g\u00f6sterim olu\u015fturulur. Son s\u0131k\u0131\u015ft\u0131rma yakla\u015f\u0131mlar\u0131, bir g\u00f6r\u00fcnt\u00fcn\u00fcn k\u00fc\u00e7\u00fck bloklar\u0131n\u0131 birbirine en \u00e7ok benzeyen (ili\u015fkili) k\u00fcmeler halinde gruplayarak ilk g\u00f6r\u00fcnt\u00fc \u00f6zelli\u011finden yararlan\u0131r.<\/span><\/p>\n

Bu \u015fekilde, her bir k\u00fcme, di\u011fer k\u00fcmelere kar\u015f\u0131l\u0131k gelenlerden ayr\u0131, etkin bir temsili ile tan\u0131mlanabilir. Genel olarak, bu, standart s\u0131k\u0131\u015ft\u0131rma \u015femalar\u0131n\u0131n yakla\u015f\u0131mlar\u0131n\u0131n sa\u011flad\u0131\u011f\u0131ndan daha verimli bir g\u00f6r\u00fcnt\u00fc g\u00f6sterimi ile sonu\u00e7lan\u0131r.<\/span><\/p>\n

\u00d6l\u00e7ekler Aras\u0131 Benzerlik<\/span><\/strong><\/h3>\n

D\u00fc\u015f\u00fck ve y\u00fcksek \u00e7\u00f6z\u00fcn\u00fcrl\u00fckl\u00fc g\u00f6r\u00fcnt\u00fclerin homolog ve y\u00fcksek oranda ili\u015fkili b\u00f6lgeleri aras\u0131nda g\u00fc\u00e7l\u00fc bir benzerlik varsa, o zaman basit bir d\u00f6n\u00fc\u015f\u00fcm\u00fcn bu kom\u015fuluklar\u0131 \u00f6l\u00e7ekler aras\u0131nda ili\u015fkilendirebilece\u011fi \u00f6ng\u00f6r\u00fclebilir. Deneysel analiz, bu t\u00fcr benzer bilgilerin \u00f6l\u00e7ekler aras\u0131nda yerel olarak var oldu\u011funu g\u00f6stermi\u015ftir; bu benzerlik, \u00f6l\u00e7ek ba\u011f\u0131ml\u0131l\u0131\u011f\u0131 olarak adland\u0131r\u0131yoruz.<\/span><\/p>\n

Bu t\u00fcr bir \u00f6l\u00e7ek kar\u015f\u0131l\u0131kl\u0131 ba\u011f\u0131ml\u0131l\u0131\u011f\u0131n\u0131n varl\u0131\u011f\u0131n\u0131n test edilmesinde deneyler yap\u0131lm\u0131\u015ft\u0131r. Deneyler, analiz edilen d\u00fc\u015f\u00fck ve y\u00fcksek \u00e7\u00f6z\u00fcn\u00fcrl\u00fckl\u00fc benzer g\u00f6r\u00fcnt\u00fclerden benzer \u015fekilde k\u00fcmelenmi\u015f homolog mahallelerin y\u00fczdesini rapor ediyor. Bu mahallelerin y\u00fcksek bir y\u00fczdesi bulunursa, mahalleler aras\u0131nda g\u00fc\u00e7l\u00fc bir \u00f6l\u00e7ekte kar\u015f\u0131l\u0131kl\u0131 ba\u011f\u0131ml\u0131l\u0131\u011f\u0131n var oldu\u011fu s\u00f6ylenir. Bu basit deneyin ayr\u0131nt\u0131l\u0131 bir a\u00e7\u0131klamas\u0131 ve analizi a\u015fa\u011f\u0131dad\u0131r.<\/span><\/p>\n

K\u0131saca, deney bir g\u00f6r\u00fcnt\u00fcn\u00fcn d\u00fc\u015f\u00fck ve y\u00fcksek \u00e7\u00f6z\u00fcn\u00fcrl\u00fckl\u00fc versiyonunu, s\u0131ras\u0131yla xl [n1 , n2 ] ve xh[n1,n2] olarak ele al\u0131r. xl[n1,n2] ve xh[n1, n2]’nin homolog yap\u0131sal kom\u015fuluklar\u0131 daha sonra K ayr\u0131k grup olu\u015fturmak i\u00e7in vekt\u00f6r niceleme (VQ) [21] kullan\u0131larak k\u00fcmelenir. K gruplar\u0131 i\u00e7in en olas\u0131 s\u0131ralaman\u0131n arand\u0131\u011f\u0131 bir kar\u0131\u015f\u0131kl\u0131k matrisi olu\u015fturulur.<\/span><\/p>\n

Son olarak, elde edilen en olas\u0131 s\u0131ralama i\u00e7in benzer \u015fekilde k\u00fcmelenmi\u015f mahallelerin y\u00fczdesinden \u00f6l\u00e7ekler aras\u0131 benzerli\u011fin bir \u00f6l\u00e7\u00fcs\u00fc elde edilir. Homolog kom\u015fulu\u011fun tan\u0131m\u0131 k\u0131saca belirtilecektir.<\/span><\/p>\n

\n

Mikrofon test<\/span>
\nWebcam test<\/span>
\nKamera test<\/span>
\nCamera<\/a> test<\/span>
\nMic test<\/span>
\nCamera online<\/span>
\nWebcam a\u00e7<\/span>
\nWebcam Driver<\/span><\/p>\n

\n

Ayr\u0131ca, yap\u0131sal bir kom\u015fulu\u011fun tan\u0131m\u0131 ve bu mahallelerin neden se\u00e7ildi\u011fi de verilecektir. \u015eimdilik, yap\u0131sal bir kom\u015fulu\u011fun bir g\u00f6r\u00fcnt\u00fc kom\u015fulu\u011funun afin e\u015flemeli bir versiyonu oldu\u011funa dikkat edin. N1 \u00d7 N2 g\u00f6r\u00fcnt\u00fcs\u00fc xl [n1 , n2 ] i\u00e7indeki H1 \u00d7 H2 kom\u015fuluklar\u0131, kom\u015fuluklar k\u00fcmesini olu\u015fturur.<\/span><\/p>\n

xl [n1 , n2 ]’nin, xh [n1 , n2 ]’den G1 \u00d7 G2 fakt\u00f6r\u00fc ile k\u0131rp\u0131larak sim\u00fcle edildi\u011fini hat\u0131rlay\u0131n. \u00d6l\u00e7ekler aras\u0131 mahalleleri sim\u00fcle etme tarz\u0131m\u0131z, sahnenin ayn\u0131 fiziksel b\u00f6lgesini kapsayan destek b\u00f6lgeleri verir, d\u00fc\u015f\u00fck \u00e7\u00f6z\u00fcn\u00fcrl\u00fckl\u00fc kom\u015fuluk, bu b\u00f6lgeyi tan\u0131mlamak i\u00e7in daha az \u00f6rne\u011fe sahiptir. Kar\u015f\u0131l\u0131k gelen g\u00f6r\u00fcnt\u00fc elde etme modeli sunulacakt\u0131r.<\/span><\/p>\n

\u015eimdi X’teki mahalleler, K ayr\u0131k grup X1, olu\u015fturmak \u00fczere k\u00fcmelenmi\u015ftir. . . , XK ve D’deki kom\u015fular, K ayr\u0131k grup D1, olu\u015fturmak i\u00e7in ayr\u0131 ayr\u0131 k\u00fcmelenir. . . , D.K. X ve D’deki homolog kom\u015fuluklar kendi g\u00f6r\u00fcnt\u00fclerinde benzer k\u00fcmeler olu\u015fturuyorsa, o zaman d\u00fc\u015f\u00fck ve y\u00fcksek \u00e7\u00f6z\u00fcn\u00fcrl\u00fckl\u00fc g\u00f6r\u00fcnt\u00fclerin bilgi i\u00e7eri\u011fi bir anlamda benzer olmal\u0131d\u0131r. Ayn\u0131 g\u00f6r\u00fcnt\u00fcden k\u00fcmelenmi\u015f bilgilerin \u00f6l\u00e7ekler aras\u0131nda ne kadar iyi ili\u015fkili oldu\u011funu belirlemek i\u00e7in g\u00f6sterildi\u011fi gibi bir kar\u0131\u015f\u0131kl\u0131k matrisi olu\u015fturuyoruz.<\/span><\/p>\n

Matrisin (j, k) konumundaki giri\u015f, Xj ve Dk,j,k = 1,…,K k\u00fcmesine atanan kom\u015fuluklar\u0131n say\u0131s\u0131d\u0131r. \u00d6l\u00e7ekler aras\u0131 kar\u015f\u0131l\u0131kl\u0131 ba\u011f\u0131ml\u0131l\u0131k, olu\u015fturulan k\u00fcmeler i\u00e7in ortak olan maksimum homolog kom\u015fuluk say\u0131s\u0131 olarak belirlenir.<\/span><\/p>\n

Xj ve Dk aras\u0131ndaki ger\u00e7ek k\u00fcmelerin veya “s\u0131n\u0131flar\u0131n” s\u0131ralamas\u0131 bilinmedi\u011finden, sadece kar\u0131\u015f\u0131kl\u0131k matrisinin k\u00f6\u015fegeninin i\u00e7eri\u011fini inceleyemeyiz. Bunun yerine, en olas\u0131 s\u0131ralamay\u0131 aramal\u0131y\u0131z. Bu da, d\u00fc\u015f\u00fck ve y\u00fcksek \u00e7\u00f6z\u00fcn\u00fcrl\u00fckl\u00fc g\u00f6r\u00fcnt\u00fclerdeki benzer bilgilerin nas\u0131l k\u00fcmelendi\u011fini g\u00f6steren bir say\u0131 verir. Bu say\u0131 a\u015fa\u011f\u0131daki basit algoritma ile kolayca bulunur.<\/span><\/p>\n

Bu y\u00fczde, k\u00fcme say\u0131s\u0131n\u0131n bir fonksiyonu olarak \u00e7izilir. K\u00fcmelenmi\u015f y\u00fcksek \u00e7\u00f6z\u00fcn\u00fcrl\u00fckl\u00fc g\u00f6r\u00fcnt\u00fc 256 \u00d7 256’d\u0131r ve resmedilmi\u015ftir. \u0130ki d\u00fc\u015f\u00fck \u00e7\u00f6z\u00fcn\u00fcrl\u00fckl\u00fc muadil g\u00f6r\u00fcnt\u00fc kullan\u0131lm\u0131\u015ft\u0131r. Biri, y\u00fcksek \u00e7\u00f6z\u00fcn\u00fcrl\u00fckl\u00fc g\u00f6r\u00fcnt\u00fcn\u00fcn (G1 = 2)\u00d7(G2 = 2) k\u0131rp\u0131lmas\u0131yla elde edildi. Di\u011fer g\u00f6r\u00fcnt\u00fc, d\u00fc\u015f\u00fck \u00e7\u00f6z\u00fcn\u00fcrl\u00fckl\u00fc g\u00f6r\u00fcnt\u00fcn\u00fcn DCT s\u0131k\u0131\u015ft\u0131r\u0131lm\u0131\u015f s\u00fcr\u00fcm\u00fcn\u00fc kullan\u0131r. Grafikler ayr\u0131ca test edilen iki farkl\u0131 kom\u015fuluk boyutunu da rapor eder: H1 \u00d7H2 =3\u00d73 ve H1 \u00d7H2 =5\u00d75.<\/span><\/p>\n

K\u00fcme say\u0131s\u0131 K = 30’a do\u011fru artt\u0131\u011f\u0131nda bile a’n\u0131n d\u00fc\u015f\u00fck \u00e7\u00f6z\u00fcn\u00fcrl\u00fckl\u00fc (s\u0131k\u0131\u015ft\u0131r\u0131lmam\u0131\u015f) g\u00f6r\u00fcnt\u00fcs\u00fc kullan\u0131ld\u0131\u011f\u0131nda, g\u00f6r\u00fcnt\u00fc \u00f6l\u00e7ekleri aras\u0131nda homolog kom\u015fuluklar\u0131n \u00e7ok g\u00fc\u00e7l\u00fc bir kar\u015f\u0131l\u0131kl\u0131 ba\u011f\u0131ml\u0131l\u0131\u011f\u0131 oldu\u011funu g\u00f6sterir.\u00a0<\/span><\/p>\n

S\u0131k\u0131\u015ft\u0131r\u0131lm\u0131\u015f d\u00fc\u015f\u00fck \u00e7\u00f6z\u00fcn\u00fcrl\u00fckl\u00fc g\u00f6r\u00fcnt\u00fc kullan\u0131ld\u0131\u011f\u0131nda kar\u015f\u0131l\u0131kl\u0131 ba\u011f\u0131ml\u0131l\u0131k azal\u0131r. Bu, \u00e7izimdeki alttaki iki e\u011fri ile g\u00f6sterilir. K = 1 durumunun her zaman 1 kar\u015f\u0131l\u0131kl\u0131 ba\u011f\u0131ml\u0131l\u0131\u011f\u0131 verdi\u011fine dikkat edin. Bunun nedeni, K = 1 i\u00e7in asl\u0131nda hi\u00e7bir k\u00fcmelemenin ger\u00e7ekle\u015ftirilmemesidir. Yani, t\u00fcm mahallelerin ayn\u0131 k\u00fcmeye ait oldu\u011fu varsay\u0131l\u0131r. Bu nedenle, “ayr\u0131k” k\u00fcmeler (veya bu durumda tek k\u00fcme) ortak olarak t\u00fcm homolog kom\u015fuluklara sahiptir.<\/span><\/p>\n

Ba\u011f\u0131ml\u0131l\u0131k genellikle k\u00fcme say\u0131s\u0131 artt\u0131k\u00e7a azal\u0131r. Bu sezgisel olarak beklenir, \u00e7\u00fcnk\u00fc k\u00fcme say\u0131s\u0131ndaki art\u0131\u015f, belirli bir g\u00f6r\u00fcnt\u00fc ve \u00f6l\u00e7e\u011fe \u00f6zg\u00fc bilgi k\u00fcmelenmesiyle sonu\u00e7lan\u0131r.<\/span><\/p>\n

D\u00fc\u015f\u00fck ve y\u00fcksek \u00e7\u00f6z\u00fcn\u00fcrl\u00fckl\u00fc benzer g\u00f6r\u00fcnt\u00fcler aras\u0131ndaki frekans i\u00e7eri\u011fi farkl\u0131 oldu\u011fundan, bir g\u00f6r\u00fcnt\u00fcdeki bilgilerin daha fazla uzmanla\u015fmas\u0131n\u0131n aralar\u0131nda daha az kar\u015f\u0131l\u0131kl\u0131 ba\u011f\u0131ml\u0131l\u0131kla sonu\u00e7lanmas\u0131 beklenir.<\/span><\/p>\n

\n\n
<\/div>\n\n<\/div>\n \n