中文摘要：

目的 目前模糊测量方法难以处理存在纹理平坦区域时的局部模糊测量.针对该问题,提出一种基于BP（back propagation）神经网络的图像局部模糊检测方法.方法 该方法采用所有奇异值以描述不同尺度信息随模糊变化情况,并与描述高频信息变化的DCT（discrete cosine transform）非零系数结合,实现奇异值向量和DCT非零系数个数联合的混合模糊测度,达到空频联合的模糊度描述.在此基础上,通过训练BP神经网络分类器,实现图像块模糊值预测.结果 单幅局部模糊图像实验中,较好区分纹理平坦区域和模糊区域的模糊程度;多幅局部模糊图像的统计实验中,召回率-准确率（RP）评估曲线显示在相同召回率下准确率较其他方法高.结论 该方法可以较准确地实现局部模糊图像（特别是存在纹理平坦区域的局部模糊图像）的模糊测量.

英文摘要：

Objective The existing blur metrics for locally blurred images are difficult to use in the measurement of flat tex- tured areas. Thus, a back propagation （BP） neural network-based image local blur measurement method is proposed to overcome this limitation. Method A new unified blur feature based on all singular values and non-zero discrete cosine transform （DCT） coefficients is presented. This feature measures sharpness from both spatial and frequency domains. Dif- ferent singular values reflect the distribution of different scale information, which vary differently after blurring. The num- ber of non-zero DCT coefficients depicts the information lost in the high frequency domain. Their combination can capture the blurring effect in the flattened textured area. BP neural network-based classifier is trained to predict the blur measure- ment of each block on the basis of the metric. Result The method can better distinguish the flat textured areas and blurred areas of a single locally blurred image compared with existing methods. According to the recall-precision curve, the statisti- cal experiment of multiple locally blurred images shows that a higher precision can be obtained with the proposed method than with existing methods. Conclusion Therefore, the proposed method measure the local blur more effectively, particular- ly that of flat textured areas, than the existing methods can.