CAJ | 학술논문

　　提出了一个包含两个自适应神经模糊推理系统和一个后处理块的网络，该网络可用于灰度图像脉冲噪声检测。网络中每个自适应神经模糊推理系统都是一个四输入单输出一阶Sugeno模糊推理系统。所提出的脉冲噪声检测方法分两步进行：对该网络进行优化训练，确定其参数；用优化后的网络对被椒盐脉冲噪声污染的图像进行噪声检测。实验结果表明，与其他传统检测方法相比，所提出的方法，更能有效检测出图像中椒盐脉冲噪声。
　　제출료일개포함량개자괄응신경모호추리계통화일개후처리괴적망락，해망락가용우회도도상맥충조성검측。망락중매개자괄응신경모호추리계통도시일개사수입단수출일계Sugeno모호추리계통。소제출적맥충조성검측방법분량보진행：대해망락진행우화훈련，학정기삼수；용우화후적망락대피초염맥충조성오염적도상진행조성검측。실험결과표명，여기타전통검측방법상비，소제출적방법，경능유효검측출도상중초염맥충조성。
A neuro-fuzzy network approach to impulse noise detection for gray scale images is presented. The network is con-structed by combining two neuro-fuzzy blocks with a postprocessor. Each neuro-fuzzy block is a first order Sugeno type fuzzy inference system with 4-inputs and 1-output. The proposed impulse noise detector consists of two modes of operation, namely, training and testing. As demonstrated by the experimental results, the proposed detector significantly outperforms other conven-tional detectors.