期刊文献+

基于SSIM_NCCDFT的超分辨率复原评价方法研究 被引量:8

Evaluation method of super-resolution restoration based on SSIM_NCCDFT
下载PDF
导出
摘要 传统的图像质量评价方法很多,但是并非针对超分辨率复原算法的特定评价指标。一种超分辨率复原算法复原性能的好坏,至今没有一个统一的评价标准,这使得超分辨率复原算法的发展受到很大限制。针对传统的超分辨率复原评价体系只关注图像某一方面统计特性的问题,提出一种基于SSIM_NCCDFT的超分辨率复原评价方法。该评价方法结合了空间域的灰度均值、对比度以及频域自相关,能够同时评价超分辨率复原结果在空间域的复原效果和对频率域信息的复原精度。实验结果表明:SSIM_NCCDFT可以准确反映图像退化的程度。相对于PSNR,SSIM_NCCDFT的优势是其同时反映了频域和空域复原的精度,评价更加全面。本文提出的基于SSIM_NCCDFT的超分辨率复原评价方法同时考虑了超分辨率复原中频率域和空间域的复原性能,评价结果较为全面,能够有效地评价复原图像中的噪声和模糊等现象,对超分辨率复原方法的评价具有一定的指导意义。 There were many traditional methods of image quality assessment.But they were not specific evaluation method for super-resolution restoration.So far there is no uniform evaluation criteria for super-resolution restoration,which largely restricted the development of super-resolution restoration algorithm.For the disadvantage of the traditional super-resolution restoration evaluation system only concerning about a particular aspect of the statistical properties of the image,this paper proposed the super-resolution restoration evaluation method based on SSIM_NCCDFT,which combined the grayvalue and contrast of the spatial domain and the autocorrelation of frequency domain.Therefore,the proposed evaluation method can evaluate the results of the super-resolution restoration in both spatial domain and frequency domain.Experimental results showed that SSIM_NCCDFT can accurately reflect the degree of image degradation.Relative to PSNR,SSIM_NCCDFT reflected both frequency and spatial accuracy.SSIM_NCCDFT was a more comprehensive evaluation.The evaluation method of super-restoration method based on SSIM_NCCDFT proposed in this paper can evaluate the results of the super-resolution restoration in both spatial domain and frequency domain.It got more comprehensive evaluation results.It can evaluate the image noise and blur.Furthermore this evaluation method has some significance for super-resolution restoration evaluation.
出处 《液晶与显示》 CAS CSCD 北大核心 2015年第4期713-721,共9页 Chinese Journal of Liquid Crystals and Displays
基金 吉林省重大科技攻关专项(No.20126015)
关键词 结构相似度 归一化傅立叶互相关系数 峰值信噪比 超分辨率复原 图像评价 SSIM NCCDFT PSNR super-resolution restoration image evaluation
  • 相关文献

参考文献15

二级参考文献284

共引文献457

同被引文献66

  • 1钟坚.国外光电侦察告警技术的装备与发展[J].舰船电子工程,2008,28(4):30-33. 被引量:5
  • 2周维虎,韩晓泉,吕大旻,崔惠绒,刘大大.军用光电系统总体技术研究[J].红外与激光工程,2006,35(z1):9-14. 被引量:10
  • 3卜莎莎,章毓晋.基于局部约束线性编码的单帧和多帧图像超分辨率重建[J].吉林大学学报(工学版),2013,43(S1):365-370. 被引量:3
  • 4刘卫光,崔江涛,周利华.插值和相位相关的图像亚像素配准方法[J].计算机辅助设计与图形学学报,2005,17(6):1273-1277. 被引量:29
  • 5YANG J C,WANG Z W,LIN Z,et al..Coupleddictionary training for image super-resolution [J].IEEE Transactions on Image Processing,2012,21(9):3467-3479.
  • 6REN C,HE X H,TENG Q Z,et al..Single imagesuper-resolution using local geometric duality andnon-local similarity[J].IEEE Transactions on Im-age Processing,2016,25(5):2169-2193.
  • 7ZHANG K B,GAO X B,LI J,et al..Single imagesuper-resolution using regularization of non-localsteering kernel regression [J].Signal Processing,2016,123:53-63.
  • 8LI J M,QU Y Y,LI C H,et al..Image super-resolu-tion based on multi-kernel regression [J].MultimediaTools and Applications,2016,75:4115-4129.
  • 9NI K S,NGUYEN T Q.Image super-resolution usingsupport vector regression [J].IEEE Transactions onImage Processing,2007,16(6):1596-1610.
  • 10YANG M C,WANG Y C F.A self-learning ap-proach to single image super-resolution[J].IEEETransactions on Multimedia, 2013,15 (3):499-509.

引证文献8

二级引证文献66

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部