摘要
针对已有彩色滤波阵列插值方法的结果图像存在边缘模糊、虚假色的问题,提出了图像相关性与支持向量机将结合的插值方法。该方法以色彩相关性为基础构造色差平面,在色差平面上根据空间相关性选择适当的邻近点输入模式训练支持向量机,然后用训练的支持向量机及输入模式估计出未知像素点对应的色差,最后计算出各像素点未知的彩色像素值。实验结果表明,与已有算法相比,该算法结果图像的PSNR值、NCD值及视觉效果均有显著改善。
To effectively reduce color artifacts and blurring of the CFA interpolation images,a support vector machines （SVM） based interpolation scheme is proposed,in which support vector regression （SVR） is used to estimate the color difference between the two color channels with applying spectral correlation of the R,G,B channels.The neighbor training sample models are selected on the color difference plane with considering spatial correlation,and the unknown color difference between two color channels is estimated by the trained SVM and input pattern,then the missing color values at each pixel can be obtained.Simulation results indicate that the proposed scheme produces visually pleasing full-color images and obtains higher PSNR and smaller NCD results than other conventional CFA interpolation algorithms.
出处
《四川大学学报（工程科学版）》
EI
CAS
CSCD
北大核心
2010年第3期145-150,共6页
Journal of Sichuan University (Engineering Science Edition)
基金
国家自然科学基金资助项目(30800240)
山东省自然科学基金资助项目(Z2007G07
Y2007G45)
哈尔滨工业大学(威海)研究基金资助项目(HIT(WH)200723)
HIT(WH)ZB200802)
关键词
图像插值
彩色滤波阵列
支持向量机
支持向量回归机
image interpolation
color filter array
Support Vector Machine （SVM）
Support Vector Regression （SVR）
