中文摘要：

针对传统无参考模糊图像质量评价方法通常需要进行学习训练或构造参考图像再进行质量评价从而导致算法较复杂且计算量大的问题，文中提出一种简捷有效的图像高频奇异值分解的无参考模糊图像质量评价方法。该方法根据自然图像同尺度不同方向高频子带小波系数相关性随模糊程度加深而降低的特性，利用奇异值分解获取图像高频子带结构特征，计算同尺度不同方向高频子带结构特征向量夹角作为质量评价指标。通过LIVE2，CSIQ和TID2013图像数据库实验表明，提出的方法与主观评价具有较好的一致性，而且算法无需训练或构造参考图像，较传统评价算法运行更为简便，实用性更强。

英文摘要：

Traditional no reference blur image quality assessment methods usually need a pre-training and learning or a reference image constructing procedure,this result in the algorithm with high computa-tion cost.Aiming to this,a simple and effective no reference blur image quality assessment algorithm is proposed based on wavelet high frequency coefficients singular value decomposition.The method is build on the observations that the different wavelet high frequency sub-bands in the same scale of an image are highly structural correlation,and the degree of correlation would be reduced as the blur distortion deepe-ning.According to this,the new method first makes wavelet transform to the image,then makes singular value decomposition to the high frequency sub-bands to get their structure information.Finally,the an-gles,which represents the similarity,between different high frequency sub-bands’structural vectors are calculated and the sum of angles is used as the last objective assessment index.Experiments results show its good effectiveness and performance on LIVE2,CSIQ and TID2013databases and compared to the tra-ditional no-reference methods,the proposed algorithm is more efficient and practical as it does not need to train or create a reference image.