Many denoising algorithms have been proposed to recover a noise corrupted image. However, most of them cannot well recover a heavy noise corrupted image with noise density above 70%. In this Letter, we propose a new approach to efficiently remove background noise by detecting and modifying noisy pixels in an image. If the center pixel of a local window is classified to noisy, this center pixel is replaced by a weighted median value on an optimum direction, enabling impulse noise to be removed. Conversely, the center pixel is kept unchanged when it is classified to noise-free, yielding the quality of restored image being well maintained. Experimental results show that the proposed approach cannot only efficiently suppress high-density impulse noise, but also can well preserve the detailed information of an image. âº Propose a modified directional-median filter to efficiently remove impulse noise. âº The quality of restored image can be well maintained. âº Proposed approach can efficiently suppress high-density (greater than 80%) noise.