基于梯度信息与HVS滤波器的无参考清晰度评价算法.doc

中图法分类号：TN911.73 文献标识码：A 文章编号： 论文引用格式: 融合梯度信息与HVS滤波器的无参考清晰度评价算法 应凌楷，张聪聪 310018 摘 要 ：目的：目前无参考图像质量评价算法的性能存在较大的提升空间，为了清晰度评价技术HVS滤波器的无参考清晰度评价算法（GI-F）。方法：该算法首先利用梯度算子计算各像素点的梯度信息，再通过HVS滤波器得到加权和作为图像的清晰度指标。结果：在公开数据库LIVE、TID2008和CSIQ上的实验结果表明,GI-F与S3（Spectral and Spatial Sharpness）、CPBD（Cumulative Probability of Blur Detection）和LPC-SI（Local Phase Coherence-based Sharpness Index）相比，性能指标RMSE（Root Mean Squared Error）、PLCC（Pearson Linear Correlation Coefficient）和SROCC（Spearman Rank-Order Correlation Coefficient）分别提升了20.66%、4.61%和3.33%；同时还的相比了%。结论：该算法只需耗费更少的时间即可计算出与人眼感知更加接近的客观清晰度指标，可广泛用于无参考图像情况下的清晰度指标计算场合，同时还可以通过 关键词 ：图像质量评价；无参考清晰度评价；梯度信息；人眼视觉系统；高通滤波器自动聚焦 A No-reference sharpness assessment algorithm with fusion of Gradient information and HVS filter Ying Lingkai College of Optical and Electronic Technology, China Jiliang University, Hangzhou 310018, China Abstract ：Objective: Distortion in digital images, inflicting suffering on the vision experience, is potentially everywhere. To some extent, distorted image can influence further research, such as the analyzing and understanding of images. In addition, the method of calculating the sharpness of image is essential for the implementation of autofocus. Indubitably, exploring the mechanism underneath the surface can perfect life progressively. Currently, there is room for improvement when it comes to the performance of no-reference image quality assessment (IQA). To upgrade the technology of sharpness assessment, an algorithm which is called GI-F and based on gradient information and HVS filter is proposed. Method: What has been demonstrated and widely accepted is that the human vision system (HVS) is highly sensitive to gradient information. In the algorithm proposed, the gradient information which is of familiarity is first calculated via the gradient operator that researchers in an accumulating number apply for the computing of image quality. Human studies in neurology have also contributed to the development of other discip