Robust estimation based image denoising with patch based dictionary

This paper proposes a method for removing fixed value impulse noise using robust estimation based filter. The proposed filter detects the outlier pixels and restore the original value using robust estimation. Initially, an adaptive dictionary is created, in which patches with similar photometric structures are clustered together. Dictionary can be constructed with the corrupted image as well as high resolution images belonging to different scenes. The denoised image is obtained from the dictionary which is trained by the K-SVD algorithm. Finally, robust estimation based filter is used to obtain the restored image. The quantitative and visual results reveals that the proposed filter outperforms the existing filters at noise level as high as 90% and also it preserves the edges and fine details effectively.