田启川, 田茂新. 一种利用神经网络提取运动模糊区域的新方法[J]. 微电子学与计算机, 2012, 29(5): 69-72.
引用本文: 田启川, 田茂新. 一种利用神经网络提取运动模糊区域的新方法[J]. 微电子学与计算机, 2012, 29(5): 69-72.
TIAN Qi-chuan, TIAN Mao-xin. A New Method for Motion-Blur Region Extraction Using Neural Network[J]. Microelectronics & Computer, 2012, 29(5): 69-72.
Citation: TIAN Qi-chuan, TIAN Mao-xin. A New Method for Motion-Blur Region Extraction Using Neural Network[J]. Microelectronics & Computer, 2012, 29(5): 69-72.

一种利用神经网络提取运动模糊区域的新方法

A New Method for Motion-Blur Region Extraction Using Neural Network

  • 摘要: 针对传统提取区域方法受背景区域影响的现象, 本文提出了一种基于频谱的人工神经网络提取模糊区域的新方法.对于局部运动模糊图像, 其完全模糊区域和局部模糊区域的频谱差异较大, 以此作为依据, 利用径向基 (RBF) 神经网络其进行训练, 逐步检测并提取出局部运动模糊图像的模糊区域.实验证明, 该方法能较好地实现局部运动模糊图像模糊区域的提取.

     

    Abstract: As for traditional methods, there is a phenomenon that extracting fuzzy region affected by background area, a new method for motion-blur region extraction of artificial neural network based on spectrum was presented.As for partial motion blur image, the Fourier spectrum amplitudes of the full fuzzy region and partial fuzzy region are quite different.According to this, trained the spectrums using radial basis function (RBF) network, detect and extract the fuzzy region of the partial motion blur image.Experiment results showed that this method can well extract motion blur region.

     

/

返回文章
返回