一种基于NIQE的印刷图像无参考质量评价方法
DOI:
作者:
作者单位:

作者简介:

通讯作者:

中图分类号:

基金项目:

新闻出版总署数字印刷工程研究中心数字传播重点实验室基金资助项目(No.10-00-309-000)


No-Reference Quality Assessment Method for Printed Image Based on NIQE
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    客观评价印刷图像质量的方法未能与主观评价达到很好的一致性,印刷图像全参考质量评价方法过于依赖参考图像,不能较好地应用在实时印刷图像质量检测系统中。因此,提出一种不需要参考图像主观评价分数的NIQE质量评价模型,并将其应用在印刷图像质量的客观评价中,结合主观实验,对印刷图像质量加以整体评价。NIQE算法与主观评价的相关性分析结果表明,皮尔逊相关系数为0.910 1,斯皮尔曼秩相关系数为0.946 1,平均绝对值误差为0.016 0;相较于MSE、PSNR、SSIM和MS-SSIM 4种全参考评价算法以及无参考评价算法BIQI,NIQE算法表现出更好的预测稳定性、单调性和一致性。因此可得,NIQE算法与人眼主观质量评价有着更好的一致性,更接近人类视觉系统,能够有效地进行实时印刷图像质量评价。

    Abstract:

    The objective evaluation method of printing image quality has not achieved good agreement with the subjective evaluation, and the reference image quality assessment method was too dependent on the reference image, which could not be used in real-time printing image quality detection system. A quality evaluation model NIQE which could do without a subjective evaluation score was proposed and applied in objective evaluation of the printed image quality to evaluate the overall quality with the combination of subjective experiments. NIQE algorithm’s results on a subjective evaluation of the correlation analysis showed the CC value being 0.910 1, the SROCC value 0.946 1, the MAE value 0.016 0, compared to the four full reference algorithms MSE, PSNR, SSIM and MS-SSIM and a no-reference evaluation algorithm BIQI, the NIQE algorithm performed better in stability, monotony and consistency. Therefore, NIQE algorithm subjective quality evaluation of the human eye had a better consistency, closer to the human visual system and could effectively evaluate the real time quality of the printed image.

    参考文献
    相似文献
    引证文献
引用本文

邵 雪,曾台英,汪祖辉.一种基于NIQE的印刷图像无参考质量评价方法[J].包装学报,2016,8(4):35-39.

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:2016-03-07
  • 最后修改日期:
  • 录用日期:
  • 在线发布日期: 2016-10-26
  • 出版日期:
文章二维码