中国图象图形学报A
中國圖象圖形學報A
중국도상도형학보A
JOURNAL OF IMAGE AND GRAPHICS
2010年
1期
26-30
,共5页
图像盲恢复%全变分正则化%四阶偏微分方程%阶梯效应
圖像盲恢複%全變分正則化%四階偏微分方程%階梯效應
도상맹회복%전변분정칙화%사계편미분방정%계제효응
image blind restoration%TV regularization%fourth-order PDE%staireasing phenomenon
由于全变分(total variation,TV)正则化方法能够很好地保持边界,因而在图像去噪和恢复中得到广泛的应用,但其一个最著的缺点就是会在恢复出来的图像中产生阶梯效应.针对此问题给出了一种基于四阶偏微分方程(PDE)的盲图像恢复模型,该模型在Chan和Wong的全变分模型的基础上,用四阶范数来代替TV范数,构造了一种新的能量泛函,消除了全变分正则化方法所产生的阶梯效应.实验结果表明,该模型能取得较好的图像恢复效果.
由于全變分(total variation,TV)正則化方法能夠很好地保持邊界,因而在圖像去譟和恢複中得到廣汎的應用,但其一箇最著的缺點就是會在恢複齣來的圖像中產生階梯效應.針對此問題給齣瞭一種基于四階偏微分方程(PDE)的盲圖像恢複模型,該模型在Chan和Wong的全變分模型的基礎上,用四階範數來代替TV範數,構造瞭一種新的能量汎函,消除瞭全變分正則化方法所產生的階梯效應.實驗結果錶明,該模型能取得較好的圖像恢複效果.
유우전변분(total variation,TV)정칙화방법능구흔호지보지변계,인이재도상거조화회복중득도엄범적응용,단기일개최저적결점취시회재회복출래적도상중산생계제효응.침대차문제급출료일충기우사계편미분방정(PDE)적맹도상회복모형,해모형재Chan화Wong적전변분모형적기출상,용사계범수래대체TV범수,구조료일충신적능량범함,소제료전변분정칙화방법소산생적계제효응.실험결과표명,해모형능취득교호적도상회복효과.
Many of the current methods for image denoising and restoration utilize total variation (TV) regularization to preserve its beneficial edge properties.However a particular caveat of TV regularization is the staircasing phenomenon in recovered images.A new blind restoration model based on a fourth-order partial differential equation (PDE) is proposed for image restoration,which improved the Chan and Wong's TV model.The new method is to introduce fourth-order norm into the energy function instead of TV norm,to overcoming the staircasing in reconstructions form TV regularization.The result of experiment shows that this restoration model is more visible.
