A New Approach for Restoring Impulse Noise Corrupted Images

In term of the value of the most noise-free pixels in the local region are close to each other, a new approach for restoring impulse noise corrupted images is proposed. It uses the standard deviation to find out the proximate 5 pixels in the filter window. Subsequently, the weighted mean value and the weighted standard deviation of the 5 pixels are used to obtain thresholds. Using the thresholds we identify the noise pixels, and then restore them with the median of the 5 pixels and the weighed mean of them. Finally, experimental tests are done on evaluating restoration quality. The test results show that our algorithm has satisfying performance on image restoring comparing to those existing approaches.