| 摘要 |
<P>PROBLEM TO BE SOLVED: To provide a most suitable noise elimination method for a color image in a multi-channel representation considering characteristics of a luminance component and a color difference component. <P>SOLUTION: An original image of the luminance component is inputted and is successively decomposed into a low-frequency sub-band and a high-frequency sub-band by wavelet transform, and respective noise components are extracted. In the same manner, an original image of the color difference component is inputted and is successively decomposed into a low-frequency sub-band and a high-frequency sub-band by wavelet transform, and respective noise components are extracted. When noise components to be eliminated from the original image of the luminance component are synthesized by inverse wavelet transform of noise components of respective sub-bands, not the extracted low-frequency noise component but the extracted high-frequency noise component is mainly used. When noise components to be eliminated from the original image of the color difference component are synthesized by inverse wavelet transform in the same manner, not the extracted high-frequency noise component but the extracted low-frequency noise component is mainly used. <P>COPYRIGHT: (C)2008,JPO&INPIT |
