透過您的圖書館登入
IP:3.141.201.253
若您是本文的作者,可授權文章由華藝線上圖書館中協助推廣。

摘要


本論文的主要目的是將多層次的紋理片段,重組於單張圖片中的物體表面。我們提出了一個新的紋理合成方法,針對多層次的半透明紋理,能夠合成出效果更為逼真的紋理。整體的處理架構可以分成兩個部份:首先,我們必須將小張的多層次紋理樣本,重組成更大的紋理影像;其次,以每一個像素為單位,計算出其各自的法向量,並以法向量的方向為基準,將法向量相似的像素聚合成單一平面,作為紋理重組的基本單位。和之前方法的不同之處,是我們在重組紋理片段時,不將其視為單層的紋理,而是將重疊區域視為多層片段的堆疊,依據每一個像素的前景機率做處理,並保留每一單層片段的紋理特徵。在平面法向量的推算過程中,我們試圖重建出一個近似的物體表面結構,並將多層次紋理依據平面法向量重組在物體表面上。除此之外,我們還可以將具有不同紋理的物件作結合,利用其前景機率產生各種不同的視覺效果。

並列摘要


The aim of this work is to transfer natural layered textures onto an unknown surface shown in a single image. The process of transferring textures can thus be divided into two parts: First, we need to recover the surface normals from the input image; second, we use a small template of natural texture to synthesize a large texture map, and then try to transfer it onto the surface according to the recovered surface normals. To produce visually pleasing results of texture transfer, we introduce a novel method of texture transfer called ''texture matting'' which can deal with layered texture and preserve their transparency to represent the layered effect.

參考文獻


[3] Alexei A. Efros and William T. Freeman. Image quilting for texture synthesis and transfer. In SIGGRAPH, pages 341-346, 2001.
[4] Alexei A. Efros and Thomas K. Leung. Texture synthesis by non-parametric sampling. In ICCV, pages 1033-1038, 1999.
[5] Cheng en Guo, Song Chun Zhu, and Ying Nian Wu. Modeling visual patterns by integrating descriptive and generative methods. International Journal of Com-puter Vision, 53(1):5-29, 2003.
[7] Aaron Hertzmann, Charles E. Jacobs, Nuria Oliver, Brian Curless, and David Salesin. Image analogies. In SIGGRAPH, pages 327-340, 2001.
[9] Anat Levin, Dani Lischinski, and Yair Weiss. A closed form solution to natural image matting. In CVPR (1), pages 61-68, 2006.

延伸閱讀


國際替代計量