Description: 摘 要 提出了一种基于样本学习的人脸肖像画自动生成算法.文章采用非均匀的马尔科夫随机场模型来描述肖
像画与人脸图像之间的统计关系 ,并使用基于训练样本的非参数化的概率表示 ,在贝叶斯优化的框架下设计了迭
代采样算法 ,可以自动的从人脸图像生成特定风格的肖像画.在该方法中 ,使用非均匀的统计模型是保持肖像中人
脸结构准确性的关键.文中所提供的例子表明了该文方法的有效性-Abstract In this paper , we present a new approach for automatically generating a life2like port rait
f rom a f rontal face image. We learn the port raiture f rom a set of real artwork examples. Different f rom
previous texture synthesis and image synthesis works that assumed modeling is homogeneous , Inhomo2
geneous Markov Random Field Model is employed as the statistical model , and a non2 paramet ric sam2
pling scheme is used to capture the complex statistical characteristics of face image and corresponding
artist drawing in this paper . In our st rategy , only those pixels corresponding to a port rait point are
sampled. Such a st rategy is crucial for maintaining facial st ructure and guaranteeing coherence of por2
t rait lines. Experimental result s demonst rate the effectiveness and life 2likeness of our approach. Platform: |
Size: 282624 |
Author:alsocc |
Hits:
Description: A non-parametric method for texture synthesis proposed.
The texture synthesis process grows a new image
outward from an initial seed, one pixel at a time. A Markov
random field model is assumed, and the conditional distribution
of a pixel given all its neighbors synthesized so far is
estimated by querying the sample image and finding all similar
neighborhoods. The degree of randomness is controlled
by a single perceptually intuitive parameter-A non-parametric method for texture synthesis is proposed.
The texture synthesis process grows a new image
outward from an initial seed, one pixel at a time. A Markov
random field model is assumed, and the conditional distribution
of a pixel given all its neighbors synthesized so far is
estimated by querying the sample image and finding all similar
neighborhoods. The degree of randomness is controlled
by a single perceptually intuitive parameter Platform: |
Size: 90112 |
Author:maulik |
Hits:
Description: Most recent approaches have posed texture synthesis in a
statistical setting as a problem of sampling from a probability
distribution. Zhu et. al. [12] model texture as a Markov
Random Field and use Gibbs sampling for synthesis. Unfortunately,
Gibbs sampling is notoriously slow and in fact
it is not possible to assess when it has converged. Heeger
and Bergen [6] try to coerce a random noise image. Platform: |
Size: 107520 |
Author:maulik |
Hits:
Search - synthesis texture markov random field