Detail-preserving impulse noise removal of images using modified dynamic programming

An algorithm that removes impulse noise in images using a modified dynamic programming method is outlined. The restoration process combines two operations. In the first operation, the pixels corrupted by impulse noise are detected by a dynamic programming based searching strategy. In the second operation, the original gray levels of the noisy pixels are estimated by a linear interpolation or extrapolation approach. A separable process is used to detect the noisy pixels of the images. In order to preserve the details of the images and reduce the computational complexity of conventional dynamic programming, two modifications have been developed. One is called neighborhood searching with an adaptive figure of merit and the other is called overlapping block processing. The actual performance of the proposed technique is compared with that of the commonly used median filter by filtering noise-corrupted images. Some computational considerations are also discussed