{"id":937,"date":"2023-03-13T22:17:42","date_gmt":"2023-03-13T19:17:42","guid":{"rendered":"https:\/\/legendhomework.com\/tr\/?p=937"},"modified":"2023-03-13T22:17:42","modified_gmt":"2023-03-13T19:17:42","slug":"mevcut-yaklasim-yontemleri-multimedya-bolumu-multimedya-bolumu-odevleri-multimedya-bolumu-tez-yaptirma-multimedya-bolumu-odev-ucretleri","status":"publish","type":"post","link":"https:\/\/legendhomework.com\/tr\/mevcut-yaklasim-yontemleri-multimedya-bolumu-multimedya-bolumu-odevleri-multimedya-bolumu-tez-yaptirma-multimedya-bolumu-odev-ucretleri\/","title":{"rendered":"Mevcut Yakla\u015f\u0131m Y\u00f6ntemleri \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":"

Mevcut Yakla\u015f\u0131m Y\u00f6ntemleri<\/span><\/strong><\/h3>\n

Perspektif d\u00f6n\u00fc\u015f\u00fcm\u00fc, denklemlerdeki rasyonel fonksiyonlar\u0131 hesaplamak i\u00e7in gereken pahal\u0131 b\u00f6lmelerden ka\u00e7\u0131nmak i\u00e7in polinomlar kullan\u0131larak yakla\u015f\u0131k olarak hesaplanabilir. Do\u011frusal yakla\u015f\u0131m, en basit ve en yayg\u0131n kullan\u0131lan yakla\u015f\u0131m tekni\u011fidir.<\/span><\/p>\n

Ancak, yakla\u015f\u0131m\u0131n basitli\u011fi nedeniyle genellikle b\u00fcy\u00fck hatalara neden olur. Daha fazla do\u011fruluk elde etmek i\u00e7in, ikinci dereceden yakla\u015f\u0131m, k\u00fcbik yakla\u015f\u0131m, iki-ikinci dereceden yakla\u015f\u0131m ve iki k\u00fcbik yakla\u015f\u0131m gibi daha karma\u015f\u0131k y\u00f6ntemler \u00f6nerilmi\u015ftir.<\/span><\/p>\n

\u00d6rt\u00fc\u015fmeyi azaltmak ve yeniden \u00f6rneklemeyi basitle\u015ftirmek i\u00e7in iki ge\u00e7i\u015fli ayr\u0131labilir algoritma gibi ek y\u00f6ntemler de geli\u015ftirilmi\u015ftir. Yakla\u015f\u0131m, perspektif d\u00f6n\u00fc\u015f\u00fcm\u00fcne yakla\u015fmak i\u00e7in de kullan\u0131lm\u0131\u015f olan say\u0131sal hesaplamada iyi bilinen bir y\u00f6ntemdir.<\/span><\/p>\n

Di\u011fer y\u00f6ntemlere g\u00f6re ana avantaj\u0131, hatas\u0131n\u0131n e\u015fit olarak da\u011f\u0131lmas\u0131d\u0131r. B\u00f6ylece sonu\u00e7 g\u00f6rsel olarak ideal sonuca daha yak\u0131n g\u00f6r\u00fcn\u00fcr. Yakla\u015f\u0131m\u0131n form\u00fcl\u00fc \u015f\u00f6yledir: Tj(x)yakla\u015f\u0131m\u0131n j’nci temel i\u015flevidir, f(x)yakla\u015f\u0131klamak i\u00e7in hedef i\u015flevdir ve N, yakla\u015f\u0131m\u0131n s\u0131ras\u0131d\u0131r. \u0130kinci dereceden Chebyshev yakla\u015f\u0131m\u0131 i\u00e7in N = 2; K\u00fcbik Chebyshev yakla\u015f\u0131m\u0131 i\u00e7in N = 3.<\/span><\/p>\n

Perspektif d\u00f6n\u00fc\u015f\u00fcm\u00fcne yakla\u015fmak i\u00e7in biquadratic ve bik\u00fcbik Chebyshev y\u00f6ntemleri de \u00f6nerilmi\u015ftir. Bu y\u00f6ntemler \u00f6nce Chebyshev kontrol noktalar\u0131n\u0131 hesaplar, ard\u0131ndan polinomlar\u0131 kullanarak rasyonel fonksiyonlara yakla\u015fmak i\u00e7in transfinite interpolasyonu kullan\u0131r.<\/span><\/p>\n

Yukar\u0131daki yakla\u015f\u0131m y\u00f6ntemlerinin t\u00fcm\u00fc, orijinal rasyonel fonksiyonlar\u0131n do\u011frudan hesaplanmas\u0131ndan daha fazla \u00e7arpma ve toplama gerektirir.<\/span><\/p>\n

Chebyshev y\u00f6ntemleri gibi karma\u015f\u0131k yakla\u015f\u0131mlar i\u00e7in, ek \u00e7arpmalar ve eklemeler, b\u00f6lmeden ka\u00e7\u0131nman\u0131n faydas\u0131n\u0131 dengeler. Do\u011frusal yakla\u015f\u0131m gibi daha basit yakla\u015f\u0131mlar, daha az ek i\u015flem gerektirir, ancak genellikle d\u00fc\u015f\u00fck kalite sa\u011flar. Bu y\u00f6ntemler ayr\u0131ca her tarama sat\u0131r\u0131ndaki yakla\u015f\u0131m katsay\u0131lar\u0131n\u0131 hesaplamak i\u00e7in bir ba\u015flatma prosed\u00fcr\u00fc gerektirir. Bu, donan\u0131m y\u00fck\u00fcn\u00fc art\u0131r\u0131r.<\/span><\/p>\n

A\u015fa\u011f\u0131daki b\u00f6l\u00fcmde, perspektif d\u00f6n\u00fc\u015f\u00fcm\u00fcn\u00fc ger\u00e7ekle\u015ftirmek i\u00e7in yeni bir y\u00f6ntem \u00f6nerilmi\u015ftir. Bu yeni y\u00f6ntem, \u00e7arpma ve toplama say\u0131s\u0131n\u0131 art\u0131rmaz, basit bir ba\u015flatma prosed\u00fcr\u00fcne sahiptir ve O(N2)’den O(N)’ye b\u00f6lme say\u0131s\u0131n\u0131 azalt\u0131r.<\/span><\/p>\n

Sabit Payda Y\u00f6ntemi<\/span><\/strong><\/h3>\n

Perspektif d\u00f6n\u00fc\u015f\u00fcm\u00fcn\u00fc sabit payda \u00e7izgileri boyunca hesaplayarak, b\u00f6lme say\u0131s\u0131 piksel ba\u015f\u0131na birden, sabit payda \u00e7izgisi ba\u015f\u0131na bire d\u00fc\u015f\u00fcr\u00fcl\u00fcr.<\/span><\/p>\n

Sabit payda y\u00f6ntemi, (d \u2212 fg), (cg \u2212 a), (af \u2212 dc), (hf \u2212e), (b\u2212hc), (ec\u2212bf ), (eg\u2212dh), ( ah\u2212bg) ve (db\u2212ae). Bu katsay\u0131lar\u0131n \u00e7er\u00e7eve ba\u015f\u0131na yaln\u0131zca bir kez hesaplanmas\u0131 gerekir. Ard\u0131ndan, (\u00f6rne\u011fin – dh) ve (ah – bg) sabit payda \u00e7izgilerinin e\u011fimini m hesaplamak i\u00e7in kullan\u0131l\u0131r.<\/span><\/p>\n

D\u00f6rt olas\u0131 durum vard\u0131r: m < -1, -1 \u2264 m \u2264 0, 0 < m \u2264 1 ve 1 < m. Durum, sabit payda \u00e7izgilerinin yatay m\u0131 yoksa dikey y\u00f6nde mi taranaca\u011f\u0131n\u0131 belirler.<\/span><\/p>\n

\u015eekil 2.1, 0 < m < 1 oldu\u011fu bir durumu g\u00f6sterir. \u00c7izgilerin t\u00fcm\u00fc m = -g’\/h’ e\u011fimine sahiptir ve sabit g’x’ + h’y’ + i’ de\u011ferlerini temsil eder. Pikseller, yakla\u015f\u0131k olarak hangi sabit payda \u00e7izgisine d\u00fc\u015ft\u00fcklerini belirtmek i\u00e7in g\u00f6lgelendirilir.<\/span><\/p>\n

\u0130lk sat\u0131r i\u00e7in pikseller, orijinden ba\u015flanarak ve Bresenham Algoritmas\u0131 uygulanarak belirlenir. Bresenham Algoritmas\u0131 yaln\u0131zca art\u0131ml\u0131 tamsay\u0131 hesaplamalar\u0131 gerektirir. Sonu\u00e7, orijinden ge\u00e7en sabit payda do\u011frusu \u00fczerindeki her yatay konum i\u00e7in kar\u015f\u0131l\u0131k gelen dikey konumu listeleyen tablodur.<\/span><\/p>\n

Tabloyu sonraki giri\u015flerin fark\u0131 olarak saklayarak, tabloyu saklamak i\u00e7in gereken bit say\u0131s\u0131 g\u00f6r\u00fcnt\u00fcn\u00fcn geni\u015flik veya y\u00fcksekli\u011finden daha fazlad\u0131r.<\/span><\/p>\n

Sabit payda \u00e7izgisinin konumu belirlendikten sonra ger\u00e7ek \u00e7arp\u0131tma ger\u00e7ekle\u015ftirilir. Paydan\u0131n tersi, \u00f6nce orijini kesen sabit payda do\u011frusu i\u00e7in hesaplan\u0131r.<\/span><\/p>\n

\n

\u00f6\u011fretim strateji, y\u00f6ntem ve teknikleri<\/span>
\nE\u011fitimde yakla\u015f\u0131m Modelleri<\/span>
\n\u00f6\u011frenme-\u00f6\u011fretme yakla\u015f\u0131mlar\u0131<\/a> ve uygulama \u00f6rnekleri<\/span>
\n\u00f6\u011frenme – \u00f6\u011fretme kuram ve yakla\u015f\u0131mlar\u0131 \u00f6zet<\/span>
\n\u00d6\u011frenme ve \u00d6\u011fretim Yakla\u015f\u0131mlar\u0131 Ders Notlar\u0131<\/span>
\nAnlat\u0131m y\u00f6ntemi Nedir<\/span>
\n\u00f6\u011frenme-\u00f6\u011fretme yakla\u015f\u0131mlar\u0131 ve uygulama \u00f6rnekleri pdf<\/span>
\nEtkili \u00f6\u011frenme ve \u00f6\u011fretme Y\u00f6ntemleri<\/span><\/p>\n

\n

Yatay konum x’, 0’dan M – 1’e art\u0131r\u0131l\u0131r; burada M, g\u00f6r\u00fcnt\u00fcn\u00fcn geni\u015fli\u011fidir. Her x’ de\u011feri i\u00e7in, \u00e7izgi tablosundan y’ elde edilir. x ve y koordinatlar\u0131n\u0131n (xn ve yn) mevcut de\u011feri, a\u015fa\u011f\u0131daki denklemler kullan\u0131larak x ve y koordinatlar\u0131n\u0131n (xn\u22121 ve yn\u22121) \u00f6nceki de\u011ferlerinden hesaplan\u0131r.<\/span><\/p>\n

Sabit payda sat\u0131r\u0131ndaki her piksel i\u00e7in xn ve yn’yi hesaplamak i\u00e7in yaln\u0131zca iki toplama gerekir. Piksel ba\u015f\u0131na \u00e7arpma veya b\u00f6lme gerekmez. Bir sonraki sabit payda \u00e7izgisi, a\u015fa\u011f\u0131daki denklem kullan\u0131larak (x’,y’) = (0,1) noktas\u0131 i\u00e7in r hesaplanarak b\u00fck\u00fcl\u00fcr.<\/span><\/p>\n

r’yi hesaplamak i\u00e7in bir toplama ve bir b\u00f6lme gereklidir. Sat\u0131r tablosu yeni sat\u0131r\u0131 izlemek i\u00e7in kullan\u0131l\u0131r ve (2.13)\u2013(2.16) denklemleri yeni sat\u0131rdaki pikselleri \u00e7arp\u0131tmak i\u00e7in kullan\u0131l\u0131r. Orijinal \u00e7izginin alt\u0131ndaki her sabit payda \u00e7izgisi \u00e7arp\u0131l\u0131r, ard\u0131ndan orijinal \u00e7izginin \u00fczerindeki sabit payda \u00e7izgileri gelir.<\/span><\/p>\n

xn ve yn genellikle tamsay\u0131lar olmad\u0131\u011f\u0131ndan, orijinal hareketli grafikte (xn, yn)’ye en yak\u0131n d\u00f6rt pikseli kullanarak \u00e7arp\u0131lm\u0131\u015f pikselin de\u011ferini hesaplamak i\u00e7in \u00e7ift do\u011frusal enterpolasyon kullan\u0131l\u0131r. B\u00fck\u00fclm\u00fc\u015f piksel P, g\u00f6sterildi\u011fi gibi a\u015fa\u011f\u0131daki \u00fc\u00e7 denklem kullan\u0131larak hesaplan\u0131r.<\/span><\/p>\n

Yukar\u0131da g\u00f6sterildi\u011fi gibi, sabit payda y\u00f6ntemi, (2.4) ve (2.5) denklemlerini do\u011frudan kullanarak piksel ba\u015f\u0131na birden (x,y) hesaplamak i\u00e7in gereken b\u00f6lme say\u0131s\u0131n\u0131 sabit payda sat\u0131r\u0131 ba\u015f\u0131na bire d\u00fc\u015f\u00fcr\u00fcr.<\/span><\/p>\n

M piksel geni\u015fli\u011finde ve N piksel y\u00fcksekli\u011finde bir g\u00f6r\u00fcnt\u00fc i\u00e7in, do\u011frudan y\u00f6ntem kullan\u0131larak MN olan b\u00f6lme say\u0131s\u0131 en fazla M + N – 1’e d\u00fc\u015f\u00fcr\u00fcl\u00fcr. (x, y) hesaplamas\u0131 i\u00e7in gereken \u00e7arpma say\u0131s\u0131 8MN’den azalt\u0131l\u0131r 8(M + N \u2212 1) + 17’ye. B\u00f6lme ve \u00e7arpmalardaki ciddi azalma, sabit payda y\u00f6ntemini ger\u00e7ek zamanl\u0131 hareketli grafik kod \u00e7\u00f6zme i\u00e7in uygun hale getirir.<\/span><\/p>\n

Ek olarak, yaln\u0131zca \u00fc\u00e7 referans noktas\u0131 iletildi\u011finde geriye do\u011fru afin d\u00f6n\u00fc\u015f\u00fcm\u00fc hesaplamak i\u00e7in sabit payda y\u00f6ntemi kullan\u0131labilir. Bu durumda d\u00fczlemdeki her nokta i\u00e7in r = 1 olur. Bu nedenle, afin d\u00f6n\u00fc\u015f\u00fcm i\u00e7in hi\u00e7bir b\u00f6lme ve \u00e7er\u00e7eve ba\u015f\u0131na yaln\u0131zca 14 \u00e7arpma gereklidir.<\/span><\/p>\n

Simulasyon Sonu\u00e7lar\u0131<\/span><\/strong><\/h3>\n

\u00c7arp\u0131tma yakla\u015f\u0131mlar\u0131n\u0131n g\u00f6rsel kalitesini kar\u015f\u0131la\u015ft\u0131rmak i\u00e7in, C++’da be\u015f y\u00f6ntem uyguland\u0131: do\u011frudan \u00e7arp\u0131tma, sabit payda, ikinci dereceden, ikinci dereceden Chebyshev ve k\u00fcbik Chebyshev. Y\u00f6ntemler daha sonra bilgisayar grafikleri i\u00e7in standart bir test g\u00f6r\u00fcnt\u00fcs\u00fc olan dama tahtas\u0131 g\u00f6r\u00fcnt\u00fcs\u00fcn\u00fc \u00e7arp\u0131tmak i\u00e7in kullan\u0131ld\u0131. Dama tahtas\u0131 g\u00f6r\u00fcnt\u00fcs\u00fc kullan\u0131\u015fl\u0131d\u0131r \u00e7\u00fcnk\u00fc perspektif d\u00f6n\u00fc\u015f\u00fcm\u00fc d\u00fcz \u00e7izgileri korumal\u0131d\u0131r.<\/span><\/p>\n

Parametreler a = 1.2, b = 0, c=\u2212100,d=0,e=1.2,f =\u221220,g=\u2212.0082,veh=0 olarak ayarlan\u0131r. Sim\u00fclasyon, orijinal g\u00f6r\u00fcnt\u00fcdeki d\u00fcz \u00e7izgilerin ikinci dereceden ve ikinci dereceden Chebyshev y\u00f6ntemleriyle b\u00fcy\u00fck \u00f6l\u00e7\u00fcde k\u0131vr\u0131ld\u0131\u011f\u0131n\u0131 g\u00f6steriyor. K\u00fcbik Chebyshev y\u00f6ntemiyle hafif\u00e7e k\u0131vr\u0131l\u0131rlar. Sabit payda y\u00f6ntemi d\u00fcz \u00e7izgileri korur.<\/span><\/p>\n

\u00c7ok \u00e7e\u015fitli durumlar i\u00e7in test verileri olu\u015fturmak \u00fczere g ve h’yi {\u2212.1, \u2212.01, \u2212.001, \u2212.0001, 0, .0001, .001, .01, .1} \u00fczerinden de\u011fi\u015ftiren sim\u00fclasyonlar ger\u00e7ekle\u015ftirildi. . a ve e parametreleri 1 olarak ayarland\u0131 ve geri kalan parametreler 0 olarak ayarland\u0131.<\/span><\/p>\n

Referans olarak do\u011frudan \u00e7arp\u0131tma g\u00f6r\u00fcnt\u00fcs\u00fc kullan\u0131larak her y\u00f6ntem i\u00e7in bir hata g\u00f6r\u00fcnt\u00fcs\u00fc hesapland\u0131 ve her bir hata g\u00f6r\u00fcnt\u00fcs\u00fcnden ortalama karesel hata (MSE) hesapland\u0131. Her bir y\u00f6ntem i\u00e7in ortalama, medyan ve maksimum ortalama hata karesi de\u011ferleri g\u00f6sterilmektedir.<\/span><\/p>\n

D\u00f6rt y\u00f6ntem i\u00e7in MSE’nin bir histogram\u0131 g\u00f6sterilmektedir. MSE, logaritmik bir \u00f6l\u00e7ekte \u00e7izilir ve 1’den k\u00fc\u00e7\u00fck t\u00fcm MSE’ler 1’de \u00e7izilir. Sabit payda y\u00f6ntemi i\u00e7in sim\u00fclasyonlar\u0131n \u00fc\u00e7te biri, 1’in alt\u0131nda MSE’lere sahipti. En b\u00fcy\u00fck hata, g = 0,01 ve h = – oldu\u011fu durumda meydana geldi. 0.1. Di\u011fer \u00fc\u00e7 y\u00f6ntem, sabit payda y\u00f6nteminden \u00f6nemli \u00f6l\u00e7\u00fcde daha az do\u011fruydu.<\/span><\/p>\n

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