【摘要】 数字全息图可为用户提供更详细的3D信息；但其编码失真及特殊的性质为其质量评价带来了挑战。提出了一种联合全息平面与物平面的全参考全息图质量评价方法。首先，在全息平面上利用信息保真度准则度量了压缩全息图的退化程度。其次，在物平面上对重建的多视口参考全息图和压缩全息图提取显著性加权的梯度相似度，再利用Grünwald-Letnikov分数阶导数度量全局特征变化；结合二者计算重建的多视口参考全息图和压缩全息图的相似度。最后，将全息平面的计算结果和物平面的相似度作为联合特征，利用随机森林对以上特征进行训练，建立质量评价模型。实验结果表明，所提出方法的PLCC约为0.917 9,SROCC约为0.907 5,优于相关代表性全息图质量评价方法，与人眼主观感受契合度更高。更多还原

【Abstract】 Digital hologram provides users with more detailed 3D information; However, its encoding distortion and the unique properties pose challenges to its quality assessment. A full reference hologram quality assessment method combining holographic plane and object plane is proposed. Firstly, the degradation degree of compressed hologram is measured using information fidelity criteria on the holographic plane. Secondly, significance weighted gradient similarity is extracted from the reconstructed multi view reference hologram and compressed hologram on the object plane, and global feature changes are measured using Grünwald-Letnikov fractional derivatives; then the similarity between the reconstructed multi view reference hologram and compressed hologram is calculated based on these two features. Finally, the calculation results of the holographic plane and the similarity of the object plane are used as joint features, and random forest is used to train the above features to establish a quality assessment model. The experimental results show that the Pearson linear correlation coefficient is about 0.917 9 and the Spearman rank correlation coefficient is about 0.907 5, which outperforms the relevant representative hologram quality assessment methods and has a higher degree of fit with the subjective perception of the human eye.更多还原