Method and apparatus for removing image noise

Abstract

Provided are a method and apparatus for removing image noise. The method includes: separating an input image signal into a signal component and a noise component; converting the noise component into a decorrelated noise component that is spatiotemporally decorrelated from neighboring pixels; and generating an image signal by adding the decorrelated noise component to the signal component.

What is claimed is: 
     
         1 . A method of removing image noise, the method comprising:
 separating an input image signal into a signal component and a noise component;   converting the noise component into a decorrelated noise component that is spatially decorrelated from neighboring pixels;   generating an out image signal that is decorrelated from noise by adding the decorrelated noise component to the signal component; and   performing temporal filtering on the out image signal to remove temporal image noise.   
     
     
         2 . The method of  claim 1 , wherein the converting of the noise component into the decorrelated noise component comprises adjusting a gain of a low frequency noise component input according to a random signal and converting the low frequency noise component into a random noise component. 
     
     
         3 . The method of  claim 1 , wherein the converting of the noise component into the decorrelated noise component comprises randomly converting a chrominance noise level of a current field into a chrominance noise level of a previous field. 
     
     
         4 . The method of  claim 1 , wherein the converting of the noise component into the decorrelated noise component comprises:
 performing a first decorrelation on luminance noise;   performing a second decorrelation on chrominance noise;   converting a luminance noise component of a current field into a high frequency random noise component; and   converting a chrominance noise level of the current field into a chrominance noise level of a previous field.   
     
     
         5 . The method of  claim 1 , wherein the converting of the noise component into the decorrelated noise component comprises:
 extracting image motion information by using a difference between an image signal of a previous field and an image signal of a current field;   determining whether the input image signal is a still image by using the image motion information;   generating a random signal if it is determined that the input image signal is a still image; and   converting a chrominance noise level of the current field into a chrominance noise level of the previous field according to the random signal.   
     
     
         6 . The method of  claim 1 , wherein the performing of temporal filtering comprises
 performing motion-compensated low-pass filtering on the output image signal.   
     
     
         7 . The method of  claim 1 , wherein the performing of temporal filtering comprises performing low-pass filtering between a pixel value of a current field and a pixel value of a previous field in which motion is estimated.