• • 上一篇    下一篇

静态图像压缩评估方法测评

文义红;杨凯;李波;   

  1. 北京航空航天大学计算机学院;
  • 出版日期:2010-02-25 发布日期:2010-02-25

Performance Test for Image Quality Assessment

Wen Yihong Yang Kai Li Bo(School of Computer Science and Technology,Beihang University,Beijing 100191)   

  • Published:2010-02-25 Online:2010-02-25
  • Supported by:
    863计划、国家自然科学基金(60775018);; 武器装备预研基金项目资助

摘要: 静态图像压缩领域提出了众多图像质量客观评估方法,但对于各种客观评估方法与主观感受相符程度的研究仍有不足。为此,文章设计了一种评价客观评估方法的测评方法:按特定原则建立相对完备的测评图像库,以实际观测得到的主观分数为参考,以客观评估方法的评价分数为测评对象,通过对比同一压缩算法不同压缩倍数、不同压缩算法主观质量相近情况下各种评估方法的表现,以及依据VQEG报告的度量准则所求的度量结果,对客观评估方法的性能进行评价。同时,文章以PSNR和SSIM为例进行了测评,结果表明SSIM各项指标优于PSNR,与现有结论一致。

关键词: 静态图像, 人眼视觉系统, 图像品质, 评估方法

Abstract: There are many objective methods for quality assessment(QA) in static image compression domain,but the research on assessing QA′s concidence with human quality judgements is inadequate.In order to compare their performances,an assessment method was designed.First build test image database,then obtain objective and subjective scores,finally compute all kinds of results based on the criteria.The test results prove that SSIM is more adapted than PSNR for assessing static image quality,and also verify the accurate of the proposed method.