【摘要】：自然界的許多生物和物體都存在著一定的對(duì)稱性。這些對(duì)稱性包括旋轉(zhuǎn)對(duì)稱,鏡像對(duì)稱以及分形等。雖然絕對(duì)的對(duì)稱十分罕見(jiàn),我們?nèi)匀豢梢詫?duì)現(xiàn)實(shí)世界中的實(shí)物建模進(jìn)行對(duì)稱性分析,探索和改進(jìn)其自身的對(duì)稱性。其中,人臉就是一個(gè)典型的、滿足鏡像對(duì)稱的重要研究對(duì)象。近年來(lái),隨著諸如微軟Kinect等商用深度攝像機(jī)的出現(xiàn)和相關(guān)領(lǐng)域的研究,我們可以經(jīng)濟(jì)、高效地采集和構(gòu)建大量的人臉模型,這使得對(duì)人臉模型進(jìn)行對(duì)稱性分析顯得愈發(fā)重要。本文提出并實(shí)現(xiàn)了一套能夠魯棒地分析人臉對(duì)稱性的工具,從而能夠?qū)θ四樐Ｐ偷膶?duì)稱性進(jìn)行逐點(diǎn)對(duì)的評(píng)分以及對(duì)稱性的修復(fù)。實(shí)驗(yàn)表明,我們提出的方法相比于傳統(tǒng)的對(duì)稱性分析方法,可以在人臉模型上獲得更好的效果。
[Abstract]:There are some symmetries in many living things and objects in nature. These symmetries include rotational symmetry, mirror symmetry, fractal and so on. Although absolute symmetry is rare, we can analyze the symmetry of physical modeling in the real world and explore and improve its own symmetry. Among them, the face is a typical, satisfying mirror symmetry important research object. In recent years, with the appearance of commercial depth cameras such as Microsoft Kinect and the research in related fields, we can collect and construct a large number of face models economically and efficiently, which makes symmetry analysis of face models become more and more important. In this paper, we propose and implement a set of tools to analyze the symmetry of human face robustly, so that the symmetry of the face model can be evaluated point by point and the symmetry can be repaired. Experiments show that compared with the traditional symmetry analysis method, the proposed method can achieve better results on the face model.
【學(xué)位授予單位】：浙江大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：TP391.41

【相似文獻(xiàn)】

相關(guān)會(huì)議論文 前1條

1 趙學(xué)莊;許秀芳;王貴昌;;分子的模糊對(duì)稱性:基本原理及簡(jiǎn)單應(yīng)用[A];中國(guó)化學(xué)會(huì)第九屆全國(guó)量子化學(xué)學(xué)術(shù)會(huì)議暨慶祝徐光憲教授從教六十年論文摘要集[C];2005年

相關(guān)博士學(xué)位論文 前1條

1 唐昌新;力學(xué)中高階張量的對(duì)稱性研究[D];南昌大學(xué);2015年

相關(guān)碩士學(xué)位論文 前2條

1 王冠華;函數(shù)對(duì)稱性問(wèn)題的教學(xué)研究[D];內(nèi)蒙古師范大學(xué);2016年

2 王安娜;超定邊界條件下偏微分方程解的對(duì)稱性[D];中國(guó)海洋大學(xué);2011年

，

本文編號(hào)：2398172