摘要
以核密度方法分别建立了图像颜色与纹理特征的概率密度模型,并用贝叶斯模型估计后验概率作为像素能量。构造的新曲线区域能量泛函分别统计内部能量与外部能量之和,以最小能量曲线对应跟踪目标的曲线。通过计算梯度下降流推进曲线演化,减少曲线能量直至收敛到目标曲线。实验结果证明,所提算法能在连续视频帧中准确地提取刚体以及非刚体目标的外部轮廓。
This paper used kernel density estimate to construct the probability distributions models for color feature and texture feature.With these two models,Bayesian model calculated the posterior probabilities of the object and background pixels.Furthermore,proposed a new region energy functional to count the energy of the object and the background pixels respectively.So that minimum energy contour coincided with the object contour.At last,the gradient descent flow derived from the variational reduced the contour energy and converge to the object boundary by evolving it.Different experimental results show that the proposed algorithm can track the rigid object contour and non-rigid object contour in image sequences efficiently.
出处
《计算机应用研究》
CSCD
北大核心
2011年第5期1954-1956,1964,共4页
Application Research of Computers
基金
国家自然科学基金资助项目(50805087
60972162)
三峡大学2009年人才科研启动基金资助项目(KJ2009B044)
关键词
目标轮廓跟踪
核密度方法
贝叶斯模型
能量泛函
水平集
track object contour
kernel density estimate
Bayesian model
energy functional
level set