基于深度置信网络的随机脉冲噪声快速检测算法

A Fast Random-valued Impulse Noise Detection Algorithm Based on Deep Belief Network

为提高现有随机脉冲噪声(RVIN)检测算法的检测准确率和执行效率,该文试图从构建描述能力更强的特征矢量和训练非线性映射更为准确的预测模型两个方面入手,实现一种基于训练策略的快速RVIN检测算法。一方面,提取多个不同阶的对数绝对差值排序统计值并结合一个能够反映图像边缘特性的统计值作为刻画图块中心像素点是否为噪声的特征矢量。在计算量增加极少的情况下,显著提升了特征矢量的描述能力。另一方面,基于深度置信网络(DBN)训练RVIN预测模型(RVIN检测器)将特征矢量映射为噪声类型标签,实现了比浅层预测模型更为准确的映射。大量实验数据表明:与现有的RVIN检测算法相比,所提算法在检测准确率和执行效率两个方面都更有优势。

To improve the detection accuracy and execution efficiency of the existing Random-Valued Impulse Noise (RVIN) detectors,a fast training-based RVIN detection algorithm is implemented by constructing a more descriptive feature vector and training a detection model with more accurate nonlinear mapping.On the one hand,multiple Rank-Ordered Logarithmic absolute Deviation (ROLD) statistics are extracted and combined with a statistical value reflecting the edge characteristics in the form of feature vector to describe how RVINlike the center pixel of a patch is.The description ability of the feature vector is improved significantly while the computational complexity is just increased in small amount.On the other hand,an RVIN prediction model (RVIN detector) is obtained by training a Deep Belief Network (DBN) to map the feature vectors to noise labels,which is more accurate than the shallow prediction model.Extensive experimental results show that, compared with the existing RVIN detectors,the proposed one has better performance in terms of detection accuracy and execution efficiency.