首页>
外国专利>
IMAGE NOISE REMOVAL USING A NON-LOCAL AVERAGE FILTERING METHOD IN A FREQUENCY CONVERSION DOMAIN FOR REMOVING NOISE BY CONVERTING 2D SIGNALS INTO A FREQUENCY DOMAIN
IMAGE NOISE REMOVAL USING A NON-LOCAL AVERAGE FILTERING METHOD IN A FREQUENCY CONVERSION DOMAIN FOR REMOVING NOISE BY CONVERTING 2D SIGNALS INTO A FREQUENCY DOMAIN
展开▼
机译:在频率转换域中使用非局部平均滤波方法去除图像噪声,以将2D信号转换为频率域来去除噪声
展开▼
页面导航
摘要
著录项
相似文献
摘要
PURPOSE: Image noise removal using a non-local average filtering method in a frequency conversion domain is provided to set adaptive variables to each sub-band in the frequency conversion domain by using the non-local average filtering method, thereby improving visual quality.;CONSTITUTION: A noise image is converted into a frequency conversion domain by a frequency conversion(11). Conversion variables of each sub-band(12) are modeled with generalized Gaussian distribution(14). A bandwidth of a non-local average filter is estimated in the conversion variables by estimating local noise statistics(15). The bandwidth is applied to non-local average filtering(16). The conversion variables are converted into a noise removal image through inverse conversion(17).;COPYRIGHT KIPO 2013;[Reference numerals] (1) Hypothesis verification in each pixel; (11) Frequency conversion; (12) Sub-band 1; (14) Sub-band noise statistical property estimation; (16) Non-local average filtering; (17) Inverse conversion; (2) Local noise statistical property estimation; (3) Local adaptive bandwidth estimation; (AA) Noise image; (BB) Sub-band K; (CC) Noise removing image
展开▼