沃尔什哈达玛变换域的无参考图像质量评价

图像失真会改变图像低频成份和图像高频成份的统计信息，基于这种特性，提出了一种新颖的无参考混合失真图像质量评价方法.首先对图像进行局部沃尔什哈达玛变换，将空域图像转换为局部沃尔什哈达玛变换图；然后在局部沃尔什哈达玛变换图上进行特征提取，即分别提取反映图像低频成份的零列率项和反映图像高频成份的非零列率项的旋转不变局部二值模式统计特征；最后利用支持向量回归网络训练特征，获得特征到质量分数的映射关系模型.在两个混合失真数据库（MLIVE数据库和MDID2013数据库）上对所提出的算法进行性能验证，实验结果表明，提出的算法具有很好的主客观评价一致性，性能优于目前现有较优秀的全参考图像质量评价算法和无参考图像质量评价算法.
Generally speaking，distortion will change the statistical characteristics of low frequency and high frequency components of images.With this consideration，a novel no-reference image quality assessment algorithm was proposed to the predict the perceived quality of multiply-distorted images.First，the images were transformed into local Walsh Hadamard Transform maps by local Walsh Hadamard Transform.Then，the features (i.e.,rotation invariant local binary pattern features of zero sequency term and non-zero sequency term) were extracted on local Walsh Hadamard Transform maps to reflect low frequency and high frequency components of images.Finally，the extracted features were trained using support vector regression to form the model，which implemented mapping from the feature space to the quality scores.Comprehensive evaluations were conducted on two multiply-distorted databases (MLIVE database and MDID2013 database)，and experimental results show that the proposed method consists well with human subjective perception.Besides，the performance of algorithm is statistically superior to the existence of better full-reference image quality assessment and non-reference image quality assessment algorithms.