本文選題：色域擴(kuò)展映射 + 膚色保護(hù)　； 參考：《山東大學(xué)》2015年碩士論文

【摘要】：隨著寬色域顯示設(shè)備的迅速發(fā)展,具有高飽和度、高亮度、高清晰度的激光顯示越來越受到重視,逐漸成為顯示領(lǐng)域研究的重點(diǎn)。激光顯示色域的顏色覆蓋率與NTSC制式相比,達(dá)到了1.8倍,與PAL制式相比,達(dá)到了1.9倍以上。但由于激光色域和現(xiàn)有制式標(biāo)準(zhǔn)并不匹配,顏色信號不能直接在激光顯示設(shè)備中顯示,否則會造成傳輸信號的失真。因此,在激光顯示中對顏色進(jìn)行空間轉(zhuǎn)換和色域的擴(kuò)展映射成為必不可缺少的步驟,以充分展現(xiàn)寬色域顯示的高飽和度高亮度的優(yōu)勢,因而研究出一組適用的色域擴(kuò)展映射算法顯得尤為重要。隨著色域擴(kuò)展映射研究的進(jìn)一步深入,在擴(kuò)展映射過程中很容易忽略膚色的處理,這就容易產(chǎn)生過度渲染和膚色與發(fā)色混亂等問題,毀壞了原有圖像的膚色分布。本文針對這一問題,主要研究了色域擴(kuò)展映射過程中的膚色保護(hù)算法。論文首先對色域擴(kuò)展映射的基本理論包括原理和研究方法進(jìn)行了詳細(xì)介紹,其次介紹了現(xiàn)有的膚色模型,并對其進(jìn)行比較對比,選出本文所用的模型,在此基礎(chǔ)之上,深入研究了色域擴(kuò)展映射過程中的膚色保護(hù)算法。本文提出的兩種色域擴(kuò)展映射過程中對膚色進(jìn)行保護(hù)的算法分別是：變色調(diào)膚色保護(hù)算法和基于概率模型的膚色保護(hù)算法。變色調(diào)膚色保護(hù)方法是針對了不同的色調(diào),將現(xiàn)有的擴(kuò)展映射算法結(jié)合起來,在拉伸的過程中,通過調(diào)整角度使得擴(kuò)展映射在線性與非線性之間轉(zhuǎn)換,最終達(dá)到顏色變化平緩無色塊,但保證不了在膚色邊界區(qū)域處的膚色點(diǎn)值沒有變化,以及當(dāng)臉上膚色與發(fā)色接近時,無法很好的保護(hù)膚色�；诟怕誓Ｐ偷哪w色保護(hù)算法是在進(jìn)行擴(kuò)展之前先判定膚色區(qū)域,當(dāng)臉上膚色與發(fā)色接近時,求取能量最小值以確定膚色區(qū)域的邊界,對于非膚色部分進(jìn)行正常的擴(kuò)展,對于膚色部分進(jìn)行較小系數(shù)的擴(kuò)展,保證擴(kuò)展之后膚色自然。通過對大量圖片進(jìn)行色域擴(kuò)展映射實(shí)驗(yàn),表明上述膚色保護(hù)算法有效,不僅消除過度渲染問題,還能準(zhǔn)確的對膚色進(jìn)行擴(kuò)展映射,有較好的擴(kuò)展效果。
[Abstract]:With the rapid development of wide color gamut display devices, laser display with high saturation, high brightness and high definition has been paid more and more attention, and has gradually become the focus of research in the field of display. Compared with NTSC, the color coverage of laser display gamut is 1.8 times and 1.9 times higher than that of pal. However, because the laser color gamut does not match with the existing standard, the color signal can not be displayed directly in the laser display device, otherwise it will cause distortion of the transmission signal. Therefore, space conversion and color gamut mapping in laser display are indispensable steps to fully demonstrate the advantages of high saturation and high brightness in wide color gamut display. Therefore, it is very important to develop a set of suitable color gamut extension mapping algorithms. With the further study of color gamut extension mapping, it is easy to ignore the skin color processing in the process of color expansion mapping, which can easily cause problems such as excessive rendering and color confusion, and destroy the skin color distribution of the original image. In order to solve this problem, this paper mainly studies the skin color protection algorithm in the process of color gamut extension mapping. Firstly, the basic theory of color gamut extension mapping is introduced in detail, then the existing skin color model is introduced and compared, then the model used in this paper is selected. The skin color protection algorithm in the process of color gamut extension mapping is studied in detail. In this paper, we propose two algorithms to protect skin color in the process of color gamut extension mapping: changing tone skin color protection algorithm and color protection algorithm based on probabilistic model. The method of changing tone skin color is to combine the existing extended mapping algorithms for different hue. In the process of stretching, the extended mapping is transformed between linear and nonlinear by adjusting the angle. Finally, the color change is smooth, but it can not guarantee the change of the color point in the skin boundary area, and when the color of the face is close to the hair color, it can not protect the skin color very well. The probabilistic model based skin color protection algorithm is to determine the skin color area before the expansion. When the skin color is close to the hair color, the minimum energy is obtained to determine the boundary of the skin color area, and the normal expansion of the non-color part is carried out. For the skin part of the expansion of a small coefficient to ensure the expansion of natural skin color. Experiments on color gamut mapping of a large number of images show that the above skin color protection algorithm is effective, not only to eliminate the over-rendering problem, but also to accurately map the skin color, which has a good expansion effect.
【學(xué)位授予單位】：山東大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2015
【分類號】：TP391.41;TN27

【參考文獻(xiàn)】

相關(guān)期刊論文 前1條

1 Charles Poynton;彭增輝;;寬色域顯示器[J];現(xiàn)代顯示;2007年12期

相關(guān)博士學(xué)位論文 前1條

1 黃國祥;RGB顏色空間及其應(yīng)用研究[D];中南大學(xué);2002年

相關(guān)碩士學(xué)位論文 前1條

1 李義輝;激光顯示的色域擴(kuò)展研究[D];山東大學(xué);2012年

，

本文編號：2062696