Impulse Noise Filtering Based On Local Histogram Partitioning

Removing impulse noise has attracted researchers’ attentions in these years. Fading impulse noise from images has high importance in image enhancement. Hence, several methods have been proposed so far for this issue. In this paper we propose a novel LHP (local histogram partitioning) technique which processes the histogram ofcorrupted image locally to discriminate between noisy and uncorrupted pixels. The experimental results have shown that LHP outperforms other noise detection techniques and produces images with higher quality. The proposed methodworks well in a wide range of impulse noise when the noise varies from 1% up to 98%. It is also useful for removing noise from color images and for real-time applications

کلیدواژه ها:

Impulse noise detection ، median filters ، histogram processing ، histogram partitioning