西安电子科技大学综合业务数字网国家重点实验室,西安,710071
网络首发:2008-10-10,
纸质出版:2008
移动端阅览
宋锐, 贾媛, 吴成柯, 等. 自适应估计模糊参数的最大后验概率超分辨率复原算法[J]. 西安交通大学学报, 2008,42(10):1254-1258.
宋锐, 贾媛, 吴成柯, et al. A Maximum Posteriori Super-Resolution Algorithm for Adaptive Estimation of Blur Parameter[J]. 2008, 42(10): 1254-1258.
为降低最大后验概率(MAP)超分辨率图像复原算法中模糊参数调整的复杂度
减少迭代运算量
提出了一种超分辨率复原新算法.先抽取一幅低分辨率图像作为参考图像
用其余低分辨率图像估计参考图像
通过训练模糊参数使估计的均方误差最小
自适应地估计最佳模糊参数.然后根据高分辨率图像和参考图像计算多项式之间的可类比性和估计误差变化的线性相关性
将训练结果直接用于超分辨率复原.复原时先利用最佳模糊参数将全部的低分辨率图像信息融合到高分辨率初始图像中
改进了复原运算的处理流程.相对于其他MAP复原算法
新算法不需要人工调整模糊参数
并且只需3次迭代即可获得稳定解
大大减少了迭代运算量.对真实图像序列的实验结果表明
新算法更好地保持了图像细节
复原图像的纹理更清晰.
A new maximum a posteriori(MAP)super-resolution algorithm is proposed to reduce the complexity of blur parameter adjustment and the iterative computation load. One low resolution(LR)image is extracted as a reference image
and then is computed by other LR images. The best parameter is adaptively estimated through training the blur parameter such that the mean square error is minimized. The parameter is then directly applied to super resolution according to the analogy of polynomial functions and the linear correlation of the estimation errors between high resolution(HR)image and LR reference image. Reconstruction procedure is improved and an initial HR image is computed by fusing all the LR images using the parameter estimation. The algorithm does not need to manually adjust the parameters and can get the best reconstruction image with only 3 iterations. The iterative computation load is greatly reduced compared with other MAP algorithms. Experiment results to real image sequences show that the texture in reconstructed image is clearer and the image detail is preserved better than other MAP algorithms.
FARSIU S, ELAD M, MILANFAR P. Multiframe demosaicing and super-resolution of color images [J]. IEEE Trans on Image Processing, 2006, 15(1): 141-159.
SHEN Huanfeng, ZHANG Liangpei, HUANG Bo, et al. A MAP approach for joint motion estimation, segmentation, and super resolution [J]. IEEE Trans on Image Processing, 2007, 16(2): 479-490.
黄华, 孔玲莉, 齐春,等.基于凸集投影和线过程模型的超分辨率图像重建[J]. 西安交通大学学报,2003,37(10):1059-1062.
HUANG Hua, KONG Lingli, QI Chun, et al. Super-resolution image reconstruction based on projections onto convex sets and line process modeling [J]. Journal of Xi'an Jiaotong University, 2003, 37(10): 1059-1062.
SCHULTZ R R, STEVENSON R L. Extraction of high-resolution frames from video sequences [J]. IEEE Trans on Image Processing, 1996, 5(6): 996-1011.
SEGALL C A, KATSAGGELOS A K, MOLINA R, et al. Bayesian resolution enhancement of compressed video [J]. IEEE Trans on Image Processing, 2004, 13(7): 898-911.
ZOMET A, RAV-ACHA A, PELEG S. Robust super-resolution [C]∥Proceeding of IEEE Computer Society Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2001: 645-650.
FARSIU S, ROBINSON D, ELAD M, et al. Fast and robust multiframe super resolution[J]. IEEE Trans on Image Processing, 2004, 13(10): 1327-1344.
BEGIN I, FERRIE F P. Blind super-resolution using a learning-based approach [C]∥Proceeding of International Conference on Pattern Recognition. Los Alamitos, CA, USA:, IEEE Computer Society, 2004: 85-89.
van de WALLE P, SUSSTRUNK S, VETTERLL M. A frequency domain approach to registration of aliased images with application to super-resolution [J]. EURASIP Journal on Applied Signal Processing, 2006(1): 1-14.
SUBBARAO M, CHOI T, NIKZAD A. Focusing techniques [C]∥Proceedings of SPIE: The International Society for Optical Engineering. Boston, MA, USA: SPIE, 1993: 163-174.
MILANFAR P, FARSIU S, ROBINSON D. MDSP super-resolution and demosaicing datasets[DB/OL].(2004-5-4)[2007-04-15]. http:∥www.soe.ucsc.edu/~milanfar/software/sr-datasets.html.
0
浏览量
5
下载量
1
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010802024621