Efficient digital image denoising for gray scale images

We propose an algorithm for image de-noising which uses a trimmed global mean filter with rank order absolute differences to remove random-valued impulse noise, stripes, scratches and blotches. It is a two stage algorithm, in the first stage the corrupted candidate is detected using rank ordered absolute differences (ROAD). In the second stage, the corrupted pixels are replaced by the median of the uncorrupted pixels in the selected window. Trimmed global mean filter is used if the selected window contains all the pixels as noisy candidate. We used a fixed window size in both detection and filtering stages. The visual and quantitative results show that proposed filter outperforms the existing filters in restoring image which is corrupted by random valued impulse noise. The proposed algorithm also provides good results over the impulse noise image corrupted by artificially introduced scratches, blotches, and stripes without causing blurring.