基于分层贝叶斯模型的图像修复方法  

IMAGE INPAINTING METHOD BASED ON HIERARCHICAL BAYESIAN MODEL

作　　者：苏挺超 沈映珊[2] Su Tingchao;Shen Yingshan(Guangzhou Open University,Guangzhou 510091,Guangdong,China;School of Computer Science,South China Normal University,Guangzhou 510631,Guangdong,China)

机构地区：[1]广州开放大学,广东广州510091 [2]华南师范大学计算机学院,广东广州510631

出　　处：《计算机应用与软件》2023年第10期261-267,共7页Computer Applications and Software

基　　金：广州市高等学校第九批教育教学改革项目(2017F10)。

摘　　要：针对当前图像修复方法存在的运算复杂、修复后图像质量下降等问题,提出一种基于分层贝叶斯模型的图像修复方法,在高斯先验假设的基础上,利用Gaussian-Wishart分布对均值和协方差矩阵的概率分布进行建模,将问题正则化的同时提升参数估计的稳定性,采用变分贝叶斯期望最大(Variational Bayesian Expectation Maximum,VBEM)算法对模型进行求解,得到最小均方误差准则下的最优重构。结果表明,该算法在图像修复时间、峰值信噪比和结构相似度等方面相对于对比方法均有较为明显提升,具有较高的应用前景。Aimed at the problems of complex operation and poor image quality in current image inpainting methods,an image inpainting method based on hierarchical Bayesian model was proposed.On the basis of Gaussian priori hypotheses,Gaussian Wishart distribution was used to model the probability distribution of mean and covariance matrix.The problem was regularized and the stability of parameter estimation was improved.Variable Bayesian expectation maximum(VBEM) algorithm was used to solve the model,and the optimal reconstruction under the minimum mean square error criterion was obtained.The results show that compared with the contrast method,the proposed algorithm can significantly improve the image repair time,peak signal-to-noise ratio and structural similarity,and it has a high application prospect.

关 键 词：图像修复 像素缺失 分层贝叶斯 先验假设 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]